Discussion:
mali gpu reverse engineering lkcl may ignore
(too old to reply)
r***@Safe-mail.net
2017-05-21 22:06:55 UTC
Permalink
Raw Message
Because I deleted a previously email about this subject, I start a new email.
Info. Lkcl said, he is not in favor of reverse engineering a mali gpu. Because it is about 150000eu and new gpus will emerge during the reverse engineering and the outcome is uncertain.

I agree on his arguments. I do not find them strong enough. Maybe lkcl assumed it was about a mali gpu on a pc card. It was not. My question was a general question about getting a broadly known mali gpu reverse engineered. In my email I referenced the https://www.kickstarter.com/projects/1771382379/firefly-rk3399-six-core-64-bit-high-performance-pl because according to lkcl it is as fast as a duo core intel mainboard and only the mali gpu software is not source code.

There is no new arm source code computer for sale. I think one should be provided. People must be able to get one. 150000eu is a crowd funding of 30000 people, each 5eu. I would pay an extra 5eu to be able to buy a source code computer.

I do not know if 30000 people are interested or if they can agree on one board.
You cannot get the mali source code faster, if you put more people on it?

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachment
Christopher Havel
2017-05-21 22:20:09 UTC
Permalink
Raw Message
Hooooooooooooooold up there, Baba Looey! (for reference
<https://en.wikipedia.org/wiki/The_Quick_Draw_McGraw_Show>)

30k is *not* a small number! 30k is a very large number. If I had 30k
pennies in my bank account right now, instead of the 44 presently there,
I'd be a happy man right now. I'd have three hundred dollars.

My entire town is 8k people. The county within which it sits is ~42k. You
are talking two thirds of the population of a slightly rural North Carolina
county.

I hope you're a real good salesman... or that you can and do hire one. You
are talking what amounts to amassing a moderately-sized army in kickstarter
terms (I use the word "kickstarter" generically here). That is a (pardon
me) f*ckton of enthusiasm to drum up.

That said... I wish you luck. You're going to need it!
mdn
2017-05-22 01:56:18 UTC
Permalink
Raw Message
If you can get a crowdfund on a used/widely spread (and if lucky still
in production) mali gpu and put the code under gpl you'll have 5 bucks
and my voice around the people I know.
The problem is always media coverage.
You need to take examples on (ugh I have to say it) hype campaigns.

I wish you luck.
Freely
BERNARD
Post by r***@Safe-mail.net
Because I deleted a previously email about this subject, I start a new email.
Info. Lkcl said, he is not in favor of reverse engineering a mali gpu. Because it is about 150000eu and new gpus will emerge during the reverse engineering and the outcome is uncertain.
I agree on his arguments. I do not find them strong enough. Maybe lkcl assumed it was about a mali gpu on a pc card. It was not. My question was a general question about getting a broadly known mali gpu reverse engineered. In my email I referenced the https://www.kickstarter.com/projects/1771382379/firefly-rk3399-six-core-64-bit-high-performance-pl because according to lkcl it is as fast as a duo core intel mainboard and only the mali gpu software is not source code.
There is no new arm source code computer for sale. I think one should be provided. People must be able to get one. 150000eu is a crowd funding of 30000 people, each 5eu. I would pay an extra 5eu to be able to buy a source code computer.
I do not know if 30000 people are interested or if they can agree on one board.
You cannot get the mali source code faster, if you put more people on it?
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
m***@gmail.com
2017-05-22 10:05:28 UTC
Permalink
Raw Message
Post by r***@Safe-mail.net
Because I deleted a previously email about this subject, I start a new email.
Info. Lkcl said, he is not in favor of reverse engineering a mali gpu.
Because it is about 150000eu and new gpus will emerge during the reverse
engineering and the outcome is uncertain.
I agree on his arguments.
I assume you don not agree.

150000eu is a crowd funding of 30000 people, each 5eu. I would pay an
Post by r***@Safe-mail.net
extra 5eu to be able to buy a source code computer.
Th issue with revese engineering the MALI gpu's is not justs about money.
ARM ltd. actively Seeks and destroys attempts on a OSS mail driver.

So that money needed is not only going to coding but is probably also
needed for legel fees and marketing against the smear and laster campaign.

They have already made one person's life very difficult:
http://libv.livejournal.com/
Post by r***@Safe-mail.net
I do not know if 30000 people are interested or if they can agree on one board.
But freeing MALI would help a lot of devices out there. So I'd trough in
some bucks. RE'ing MALI would not be for just one board.
Post by r***@Safe-mail.net
You cannot get the mali source code faster, if you put more people on it?
Finding the right minds and right amount of them working on the same thing
is a hard equation.

You could add me to that team but my skills would be of limited use. Adding
someone of the same skill set would probably be even less effective.

So more money or more people is not the solutions. The right people and the
right amount is needed.
Post by r***@Safe-mail.net
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
mdn
2017-05-22 18:01:23 UTC
Permalink
Raw Message
Post by m***@gmail.com
Post by r***@Safe-mail.net
Because I deleted a previously email about this subject, I start a new email.
Info. Lkcl said, he is not in favor of reverse engineering a mali gpu.
Because it is about 150000eu and new gpus will emerge during the reverse
engineering and the outcome is uncertain.
I agree on his arguments.
I assume you don not agree.
150000eu is a crowd funding of 30000 people, each 5eu. I would pay an
Post by r***@Safe-mail.net
extra 5eu to be able to buy a source code computer.
Th issue with revese engineering the MALI gpu's is not justs about money.
ARM ltd. actively Seeks and destroys attempts on a OSS mail driver.
So that money needed is not only going to coding but is probably also
needed for legel fees and marketing against the smear and laster campaign.
http://libv.livejournal.com/
Ho yes I remember that now.
The thing is, is that one wants to avoid to be persecuted you have to:
-Avoid electronic money (or at least one that doesn't tie your real name
too it).
-Having your name tied to the project it (staying anonymous)
-Having people who don't have a lot to loose (if they can stay anonymous
this isn't a big problem).
-Having a server that won't reveal the ID of the people participating on
the project (Tor or GNUnet only)
There are some projects that exist already on these mesh networks.

The problem will always be money and official crowdfunding systems that
needs to have IDs (which is kinda normal).
Post by m***@gmail.com
Post by r***@Safe-mail.net
I do not know if 30000 people are interested or if they can agree on one board.
But freeing MALI would help a lot of devices out there. So I'd trough in
some bucks. RE'ing MALI would not be for just one board.
Post by r***@Safe-mail.net
You cannot get the mali source code faster, if you put more people on it?
Finding the right minds and right amount of them working on the same thing
is a hard equation.
You could add me to that team but my skills would be of limited use. Adding
someone of the same skill set would probably be even less effective.
So more money or more people is not the solutions. The right people and the
right amount is needed.
Post by r***@Safe-mail.net
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Neil Jansen
2017-05-22 21:28:21 UTC
Permalink
Raw Message
Post by mdn
-Avoid electronic money (or at least one that doesn't tie your real name
too it).
-Having your name tied to the project it (staying anonymous)
-Having people who don't have a lot to loose (if they can stay anonymous
this isn't a big problem).
-Having a server that won't reveal the ID of the people participating on
the project (Tor or GNUnet only)
There are some projects that exist already on these mesh networks.
Crowdfunding and anonymity DO NOT go together.

I'm not backing a project from some random hacker via bitcoin or Western
Union, nope not going to happen.

There are enough problems with crowdfunding campaigns that aren't
anonymous, the way I see it.


I'd crowdfund a legitimate (non-anonymous campaign) to (A) raise awareness
for the need of FOSS graphics drivers, (B) help fund a legal group to make
it easier for hackers and hobbyists to reverse engineer what they have
purchased, or (C) raise funds to help individuals like Luc Verhaegen that
have gotten smeared but want to keep fighting. Or (D) if anyone in a
matter of years/decades/whatever is crazy enough to attempt a completely
FOSS GPU core which won't come cheap or easy. Something like Project VGA
but scaled up in complexity and down in size.

But that's about it. Other people's mileage many vary. But being that
Wikipedia and other sources aren't even really mentioning all of the
litigation and smearing going on, I'd suggest to start by maybe getting the
word out, maybe edit a few wikipedia pages to make that a bit more
apparent. I don't even see it mentioned in the talk pages. This group was
literally the first time that I've heard of it.
Bill Kontos
2017-05-22 10:21:11 UTC
Permalink
Raw Message
Post by r***@Safe-mail.net
I do not know if 30000 people are interested or if they can agree on one board.
You cannot get the mali source code faster, if you put more people on it?
Look at what happened to Luc Verhaegen. Arm has destroyed his life. The
company is hard at work killing any attempt at a libre driver. You will
have a hard time finding people willing to work for this. Also 30k is a lot
of people for a kickstarter.
Post by r***@Safe-mail.net
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Luke Kenneth Casson Leighton
2017-05-26 20:41:32 UTC
Permalink
Raw Message
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
what i *also* tend to do is to tidy up the sentences adding line-breaks
as well as carriage-returns that separate out the paragraphs. this
I will shorten lines.
thx ron that's perfect.
if you would like your ideas to be easier to understand perhaps you
might consider switching to "plain text only" mode in your email client
I always write in plain text. If I do not, it is a mistake.
you're writing in plaintext... but your mailer is definitely selecting
"rich text" (i.e. HTML) to do it.

if you haven't changed anything in your mailer, the action i took
to have mailman always do a conversion from HTML to plaintext
is doing its job perfectly, so you don't have to do anything.

yay for mailman.

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@files
Luke Kenneth Casson Leighton
2017-05-25 20:48:05 UTC
Permalink
Raw Message
folks, it's very interesting to note this discussion as being very (visually)
difficult to read, because normally i reply to pretty much every discussion
and i take particular care to do inline posting and to cut unnecessary
context (as is being done here)... but also, from experience,
it would seem that i do quite a bit more than that.

what i *also* tend to do is to tidy up the sentences adding line-breaks
as well as carriage-returns that separate out the paragraphs. this
helps the people who reply to be able to not only identify the different
people in the conversation but also it helps their mailers to add in
the correct level of additional indentation - the ">"s before each
line.

now, if you *don't* do that, then you end up with an absolute mess:
one single ">" per paragraph... but you can't identify it visually
because there's *no paragraph breaks*.

ron, it is *almost impossible* to identify what you've written, thus making
it much harder to understand what you want to say, and thus leading
directly to the confusion and lack of clarity that you're experiencing.

if you would like your ideas to be easier to understand perhaps you
might consider switching to "plain text only" mode in your email client
and to re-read any online netiquette rules for email lists once again, paying
particular attention to layout and the use of whitespace in replies.

l.
-------- Original Message --------
If you reinterpret what I write then tell on what grounds. I think lkcl's reasons for not reverse engineering a mali are right.
It is difficult to follow your argumentation. You write that you agree
and disagree in the next sentence.
No, I do not.
_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large
Pablo
2017-05-24 20:40:07 UTC
Permalink
Raw Message
If you reinterpret what I write then tell on what grounds. I think lkcl's reasons for not reverse engineering a mali are right.
It is difficult to follow your argumentation. You write that you agree
and disagree in the next sentence.
Weighing them up against the importance of getting a libre software gpu, I reach the conclusion that the reverse engineering should be done.
Before you plan a difficult crowd funding campaign and involve the FSF, please tell us your
counter arguments to Lukes reasoning on a technical and ethical level.
For example to the following points I am quoting from Lukes email reply to you in
the thread about "firefly 3399 all source software disclosed?":
- "take one of the "open gpus" or parts of them and use that."
- "the sad fact of reverse-engineering: all that effort, with
*no guarantee of success*.... just to get something that's years
out-of-date."
- "well, with the same money it would be possible to make our
own libre processor, with enough extensions to be able to do 3D
graphics *without* paying anyone a cent."
There have been some remarks about the probability of a successful crowd funding. I mentioned the numbers 50000 people, each paying 5euro. Notice on this email list, people want to pay 5eu, if they get the software in question.
It is safe to say more than 10 million people have gnulinux on their computer? A major part of them know about the importance of libre software and a part of them would want to act on it.
I don't know if your numbers are correct but it seems important to me to
point out that only a fraction of all GNU/Linux Users own a device with
a mali GPU. Only a part of that group would in principle support such a
campaign. Only some of the willing will actually fund the campaign...

Pablo

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@fil
d***@mail.com
2017-05-30 20:48:36 UTC
Permalink
Raw Message
On Mon, 22 May 2017 12:05:28 +0200
Post by r***@Safe-mail.net
Because I deleted a previously email about this subject, I start a new
email.
Info. Lkcl said, he is not in favor of reverse engineering a mali gpu.
Because it is about 150000eu and new gpus will emerge during the
reverse engineering and the outcome is uncertain.
I agree on his arguments.
I assume you do not agree.
Post by r***@Safe-mail.net
150000eu is a crowd funding of 30000 people, each 5eu. I would pay an
extra 5eu to be able to buy a source code computer.
The issue with revese engineering the MALI gpu's is not justs about
money. ARM ltd. actively Seeks and destroys attempts on a OSS mail
driver.
So that money needed is not only going to coding but is probably also
needed for legel fees and marketing against the smear and laster
campaign.
http://libv.livejournal.com/
I'd do it in a heart beat in spite of the evil people out there if I had
the skills.

Incidentally, couldn't you do this in an mostly automated fashion?
I mean:
1. Have computer program send bits to GPU
2. Have same program read bits from EDP (or whatever), to determine
result and time spend on task.
4. Have another program create a spread sheet for the in-out-time info
and stats.
3. Have developer look at these and code.
Post by r***@Safe-mail.net
I do not know if 30000 people are interested or if they can agree on
one board.
But freeing MALI would help a lot of devices out there. So I'd trough in
some bucks. RE'ing MALI would not be for just one board.
Agreed. It must be for all boards and support opengl, opencl, vulkan.
Post by r***@Safe-mail.net
You cannot get the mali source code faster, if you put more people on
it?
Finding the right minds and right amount of them working on the same
thing is a hard equation.
You could add me to that team but my skills would be of limited use.
Adding someone of the same skill set would probably be even less
effective.
So more money or more people is not the solutions. The right people and
the right amount is needed.
Well, I'd put ten dollars to a campaign like this without a HW reward.
I'm assuming that beings that there are so many Mali GPUs and hacker
boards out there that other people would also be very interested in this.
What would I need to know to do this?
Quick, point me to the books!
No, really, I would do such a thing, I don't have a lot to loose, though
for free I'd be taking my time...
But still, I'd need an education.

Thanks,
David

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to a
sam via arm-netbook
2017-05-30 22:16:42 UTC
Permalink
Raw Message
Post by d***@mail.com
Well, I'd put ten dollars to a campaign like this without a HW reward.
I'm assuming that beings that there are so many Mali GPUs and hacker
boards out there that other people would also be very interested in this.
What would I need to know to do this?
Quick, point me to the books!
No, really, I would do such a thing, I don't have a lot to loose, though
for free I'd be taking my time...
But still, I'd need an education.
Start from the existing code http://limadriver.org/



_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@files
David Niklas
2017-06-17 13:39:36 UTC
Permalink
Raw Message
On Wed, 31 May 2017 00:16:42 +0200
Post by sam via arm-netbook
Post by d***@mail.com
Well, I'd put ten dollars to a campaign like this without a HW reward.
I'm assuming that beings that there are so many Mali GPUs and hacker
boards out there that other people would also be very interested in
this. What would I need to know to do this?
Quick, point me to the books!
No, really, I would do such a thing, I don't have a lot to loose,
though for free I'd be taking my time...
But still, I'd need an education.
Start from the existing code http://limadriver.org/
I understand that I understand very little about this process so please
read this with hearty laugh prepared.

Whereas other older people have gone from simple reverse engineering
projects to more difficult ones I have come into the game late when all
the projects are the most difficult.
Let me assume that the GPU is a RISC model and uses 8-bit instructions.
Then it would have a total of 255 instructions (the 256th would be all
zeros and be a no-op because the wires on the line need a way to tell if
they have an instruction on them and that is the most power conservative
I can think of).
Now let use assume that if the signed bit is set that the GPU receives an
instruction to set an internal option.
All that I could probably learn from the lima driver and also some idea
of how the 2D rendering engine works and what it's instructions are.
Now the questions come up:
1. What are the options for the 3D engine?
2. What are the instructions for the 3D engine?
3. How do the 2D, 3D, and video (de|en)code engine fit together?

To sum it up, I don't think it's as simple as downloading the code,
signing up for the mailing list, and coding. It might be, someone could
have left full specs laying around waiting to be turned into mock-up code
and then real code; but I doubt it.
That's not to say I will not try, but I just don't see this as a very
productive path.

I suffer from the black box discouragement effect. Someone builds a black
box, then a bigger black box, then an even larger black box; eventually
no one knows how it works inside, even the people who designed it
understand only a relatively small part.

Sincerely,
David

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@files.p
zap
2017-06-17 18:53:47 UTC
Permalink
Raw Message
Post by David Niklas
On Wed, 31 May 2017 00:16:42 +0200
Post by sam via arm-netbook
Post by d***@mail.com
Well, I'd put ten dollars to a campaign like this without a HW reward.
I'm assuming that beings that there are so many Mali GPUs and hacker
boards out there that other people would also be very interested in
this. What would I need to know to do this?
Quick, point me to the books!
No, really, I would do such a thing, I don't have a lot to loose,
though for free I'd be taking my time...
But still, I'd need an education.
Start from the existing code http://limadriver.org/
I understand that I understand very little about this process so please
read this with hearty laugh prepared.
Whereas other older people have gone from simple reverse engineering
projects to more difficult ones I have come into the game late when all
the projects are the most difficult.
Let me assume that the GPU is a RISC model and uses 8-bit instructions.
Then it would have a total of 255 instructions (the 256th would be all
zeros and be a no-op because the wires on the line need a way to tell if
they have an instruction on them and that is the most power conservative
I can think of).
Now let use assume that if the signed bit is set that the GPU receives an
instruction to set an internal option.
All that I could probably learn from the lima driver and also some idea
of how the 2D rendering engine works and what it's instructions are.
1. What are the options for the 3D engine?
2. What are the instructions for the 3D engine?
3. How do the 2D, 3D, and video (de|en)code engine fit together?
To sum it up, I don't think it's as simple as downloading the code,
signing up for the mailing list, and coding. It might be, someone could
have left full specs laying around waiting to be turned into mock-up code
and then real code; but I doubt it.
That's not to say I will not try, but I just don't see this as a very
productive path.
I suffer from the black box discouragement effect. Someone builds a black
box, then a bigger black box, then an even larger black box; eventually
no one knows how it works inside, even the people who designed it
understand only a relatively small part.
IF you can figure out how to reverse engineer the 3d engine, ikcl would
be very happy I am sure. But I get the feeling from him that it is
nearly impossible to do this.

Not sure why... but yeah... RK3288 with 3d engine would be cool I will
admit that. or an even later more compatible processor.

Still, I think you have an uphill battle for that.

ps, I don't have much reverse engineering experience at all, so what I
am suggesting, could be easier or harder.

I just don't know. xD
Post by David Niklas
Sincerely,
David
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large at
Luke Kenneth Casson Leighton
2017-06-18 06:09:40 UTC
Permalink
Raw Message
Post by zap
IF you can figure out how to reverse engineer the 3d engine, ikcl would
be very happy I am sure.
that used to be the case but is no longer true. please read section
2.5 of the bill of ethics before proceeding further with this thread:
https://www.titanians.org/the-bill-of-ethics/

ARM's illegal and unethical activity which destroyed luc verhaegen's
career and reputation - including several counts of slander as well as
blackmail of the company funding his reverse-engineering efforts - was
the last straw.

if we proceed to reverse-engineer MALI, logically it results in
people buying more ARM products.

if people buy more ARM products, logically it results in more money
(resources) going to support ARM's illegal and unethical actions.

any action which is taken that results in support or endorsement of
unethical actions is, logically, itself, unethical

therefore, logically and plainly put: unless ARM's attitude changes
the reverse-engineering of MALI is itself an unethical act.

so it's a simple chain of logical reasoning based on ethical principles.

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Se
Bill Kontos
2017-06-18 08:16:38 UTC
Permalink
Raw Message
Post by Luke Kenneth Casson Leighton
if we proceed to reverse-engineer MALI, logically it results in
people buying more ARM products.
if people buy more ARM products, logically it results in more money
(resources) going to support ARM's illegal and unethical actions.
any action which is taken that results in support or endorsement of
unethical actions is, logically, itself, unethical
therefore, logically and plainly put: unless ARM's attitude changes
the reverse-engineering of MALI is itself an unethical act.
So someone buying an nvidia card and using nouveau with it is by your
logic unethical ? Tell that to rms he recommends old nvidia cards :P
Besides I fail to see what else apart form more sales could convince
arm to open their drivers.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large a
Jean Flamelle
2017-06-18 08:25:50 UTC
Permalink
Raw Message
Aftermarket devices fail to support original dev's the same way used
books fail to support original authors, so that is also a
consideration.
Post by Bill Kontos
Post by Luke Kenneth Casson Leighton
if we proceed to reverse-engineer MALI, logically it results in
people buying more ARM products.
if people buy more ARM products, logically it results in more money
(resources) going to support ARM's illegal and unethical actions.
any action which is taken that results in support or endorsement of
unethical actions is, logically, itself, unethical
therefore, logically and plainly put: unless ARM's attitude changes
the reverse-engineering of MALI is itself an unethical act.
So someone buying an nvidia card and using nouveau with it is by your
logic unethical ? Tell that to rms he recommends old nvidia cards :P
Besides I fail to see what else apart form more sales could convince
arm to open their drivers.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to
Luke Kenneth Casson Leighton
2017-06-18 14:32:53 UTC
Permalink
Raw Message
Post by Bill Kontos
Post by Luke Kenneth Casson Leighton
if we proceed to reverse-engineer MALI, logically it results in
people buying more ARM products.
if people buy more ARM products, logically it results in more money
(resources) going to support ARM's illegal and unethical actions.
any action which is taken that results in support or endorsement of
unethical actions is, logically, itself, unethical
therefore, logically and plainly put: unless ARM's attitude changes
the reverse-engineering of MALI is itself an unethical act.
So someone buying an nvidia card and using nouveau with it is by your
logic unethical ?
how did you gain that impression from a logical chain of "ANDed"
statements which began with some illegal and unethical behaviour by
ARM?

i note from the above context that you cut off the absolutely
critical part of the chain of ANDed statements. you specifically cut
the one *right* at the beginning, which was the illegal and unethical
behaviour that ARM committed against luc verhaegen.

did you not read that part, or were you under the impression that it
was not part of the chain of logical reasoning?

has NVIDIA engaged in similar illegal or unethical practices to that
which ARM carried out against luc verhaegen, which we know were and
are sanctioned by the CTO of ARM?

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Sen
Bill Kontos
2017-06-18 16:03:29 UTC
Permalink
Raw Message
Post by Luke Kenneth Casson Leighton
has NVIDIA engaged in similar illegal or unethical practices to that
which ARM carried out against luc verhaegen, which we know were and
are sanctioned by the CTO of ARM?
Sorry I chopped it off to keep the conclusion only. Nvidia switched to
requiring signed binary firmware with a blown fuse for any maxwell or
newer card to run effectively blocking the work by the libre driver
and delaying the release of the signed binaries to 2 years after the
cards are actually released to the market. So they are actively
blocking reverse engineering.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to ar
zap
2017-06-18 16:56:20 UTC
Permalink
Raw Message
Post by Luke Kenneth Casson Leighton
Post by zap
IF you can figure out how to reverse engineer the 3d engine, ikcl would
be very happy I am sure.
that used to be the case but is no longer true. please read section
https://www.titanians.org/the-bill-of-ethics/
ARM's illegal and unethical activity which destroyed luc verhaegen's
career and reputation - including several counts of slander as well as
blackmail of the company funding his reverse-engineering efforts - was
the last straw.
if we proceed to reverse-engineer MALI, logically it results in
people buying more ARM products.
if people buy more ARM products, logically it results in more money
(resources) going to support ARM's illegal and unethical actions.
any action which is taken that results in support or endorsement of
unethical actions is, logically, itself, unethical
therefore, logically and plainly put: unless ARM's attitude changes
the reverse-engineering of MALI is itself an unethical act.
so it's a simple chain of logical reasoning based on ethical principles.
l.
Hmm... just out of curiosity, what is your plan then? to make your own
processors from lowrisc?
its not a bad idea, but I think until that is an option... we should use
still use some form of arm. Unless you know of other options.

Just curious but what other options are there? Also, I think that makes
it more reasonable to reverse engineer their products just to piss Arm
off. They don't like their products being reverse engineered anyways...
so why not do that to annoy them for their unethical acts? Besides it
could make them realize that their evil actions need to be changed.

Also, the developer who was blackmailed may be pleased by this course of
action.

Its not my favorite idea, but its better than letting mali run
unchecked. In my opinion.

You are of course free to disagree but that's my stance.
Post by Luke Kenneth Casson Leighton
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@files.phc
Luke Kenneth Casson Leighton
2017-06-18 19:14:09 UTC
Permalink
Raw Message
Post by zap
Hmm... just out of curiosity, what is your plan then? to make your own
processors from lowrisc?
replying in part to bill here as well: yes. and to use MIAOW for
OpenCL and ORSOC GPU for actual rendering. it won't be perfect but it
will be a start.

bill: nvidia are in the difficult position of likely having been
pressurised by governments to lock down what is effectively viewed in
military terms as a a weapon (the rest of us just call it a "GPU").
if you recall many years ago, iraq i believe it was purchased
thousands of sony PS1s to make a supercomputer.

as there is an ongoing arms race in that regard it is only the latest
processors which are likely to fall under, for example, U.S. BXPA
Weapons-Grade "Munitions" classification. given the fact that it is
after a couple of years that the source code is no longer
DRM-restricted, we have a correlation that fits with the ongoing
evidence.

now, as long as a replacement (libre) processor is well below the
"state of the art" but is otherwise perfectly acceptable for
mass-volume electronics purposes, it will fall outside of this
potential trap.
Post by zap
its not a bad idea, but I think until that is an option... we should use
still use some form of arm.
indeed. it may sound strange but when there is no other option (and
by that i mean *exhaustive* analysis finds no other option) i do not
mind "crossing the line" into what would traditionally be viewed by
software libre purists as "unacceptable territory" *IF* in doing so it
is part of a long-term strategy to *REPLACE* the very thing being
leveraged [to make money etc. etc.]

for example: many software libre supporters flatly refuse to even
*install* Windows NT... but if i had taken that attitude i would not
have broken the NT Domains protocol, over 20 years ago.
Post by zap
Unless you know of other options.
nope, i don't. always looking though.
Post by zap
Just curious but what other options are there? Also, I think that makes
it more reasonable to reverse engineer their products just to piss Arm
off. They don't like their products being reverse engineered anyways...
so why not do that to annoy them for their unethical acts? Besides it
could make them realize that their evil actions need to be changed.
and remove the one thing which would otherwise teach them a lesson?

i see both perspectives: i just believe that they are sufficiently
arrogant in their power and beliefs that it is unlikely that they will
change their minds. they've been told by their engineers countless
times. they've been told by users countless times. they've been told
by businesses who would otherwise buy more of their products countless
times.
Post by zap
Its not my favorite idea, but its better than letting mali run
unchecked. In my opinion.
yehyeh, i hear ya.
Post by zap
You are of course free to disagree but that's my stance.
no it's good to hear. thx zap.

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
zap
2017-06-18 22:02:01 UTC
Permalink
Raw Message
Post by Luke Kenneth Casson Leighton
Post by zap
Hmm... just out of curiosity, what is your plan then? to make your own
processors from lowrisc?
replying in part to bill here as well: yes. and to use MIAOW for
OpenCL and ORSOC GPU for actual rendering. it won't be perfect but it
will be a start.
bill: nvidia are in the difficult position of likely having been
pressurised by governments to lock down what is effectively viewed in
military terms as a a weapon (the rest of us just call it a "GPU").
if you recall many years ago, iraq i believe it was purchased
thousands of sony PS1s to make a supercomputer.
as there is an ongoing arms race in that regard it is only the latest
processors which are likely to fall under, for example, U.S. BXPA
Weapons-Grade "Munitions" classification. given the fact that it is
after a couple of years that the source code is no longer
DRM-restricted, we have a correlation that fits with the ongoing
evidence.
now, as long as a replacement (libre) processor is well below the
"state of the art" but is otherwise perfectly acceptable for
mass-volume electronics purposes, it will fall outside of this
potential trap.
Please use lowrisc if you do this option, they already are libre. Their
stuff is licensed under gpl3. That should also mean its easier to,
load/less risk of idiots trying to but proprietary crap into it and get
away with it like google does. bleh... google is so awful.
Post by Luke Kenneth Casson Leighton
Post by zap
its not a bad idea, but I think until that is an option... we should use
still use some form of arm.
indeed. it may sound strange but when there is no other option (and
by that i mean *exhaustive* analysis finds no other option) i do not
mind "crossing the line" into what would traditionally be viewed by
software libre purists as "unacceptable territory" *IF* in doing so it
is part of a long-term strategy to *REPLACE* the very thing being
leveraged [to make money etc. etc.]
for example: many software libre supporters flatly refuse to even
*install* Windows NT... but if i had taken that attitude i would not
have broken the NT Domains protocol, over 20 years ago.
I am glad wine was created, too bad that I cannot plan windows 95 games
through wine yet... completely I mean.
Post by Luke Kenneth Casson Leighton
Post by zap
Unless you know of other options.
nope, i don't. always looking though.
That is good.
Post by Luke Kenneth Casson Leighton
Post by zap
Just curious but what other options are there? Also, I think that makes
it more reasonable to reverse engineer their products just to piss Arm
off. They don't like their products being reverse engineered anyways...
so why not do that to annoy them for their unethical acts? Besides it
could make them realize that their evil actions need to be changed.
and remove the one thing which would otherwise teach them a lesson?
i see both perspectives: i just believe that they are sufficiently
arrogant in their power and beliefs that it is unlikely that they will
change their minds. they've been told by their engineers countless
times. they've been told by users countless times. they've been told
by businesses who would otherwise buy more of their products countless
times.
Dunno, I thought it was a good idea.
Post by Luke Kenneth Casson Leighton
Post by zap
Its not my favorite idea, but its better than letting mali run
unchecked. In my opinion.
yehyeh, i hear ya.
Post by zap
You are of course free to disagree but that's my stance.
no it's good to hear. thx zap.
l.
Tell me what you think of lowrisc when you get a chance. I mean as a
base for your processors. heh.
Post by Luke Kenneth Casson Leighton
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@files.phcomp.co
Luke Kenneth Casson Leighton
2017-06-18 22:45:09 UTC
Permalink
Raw Message
Post by zap
Tell me what you think of lowrisc when you get a chance. I mean as a
base for your processors. heh.
i feel that the lowrisc team fit better with the ethics of what i
would like to see be achieved (note deliberate use of third person,
i.e. that i am not personally tying a personal egoistic agenda to or
correlation with the desire to see success).

i noted in particular that the lowrisc team has set up as a CIC.
that's a big plus.

technically i am particularly impressed with the concept of using a
32-bit RISCV for GPIO, which they call "minion cores". bitbanging
isn't really bitbanging any more if there's an entire CPU dedicated to
it. the advantage of their approach is that you no longer require
complex multiplexing hardware on the GPIO (as is normally done, with
dedicated hardware blocks for each I/O function). you simply... load
a different program into the minioncore and the pins which e.g. were
previously I2C are now UART. or
some-other-future-as-yet-unspecified-or-unforseen-I/O-interface.
upgrading to the latest version of SDMMC is therefore dead easy: just
write a new program for the minion core.

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send
zap
2017-06-19 01:06:10 UTC
Permalink
Raw Message
Post by Luke Kenneth Casson Leighton
Post by zap
Tell me what you think of lowrisc when you get a chance. I mean as a
base for your processors. heh.
i feel that the lowrisc team fit better with the ethics of what i
would like to see be achieved (note deliberate use of third person,
i.e. that i am not personally tying a personal egoistic agenda to or
correlation with the desire to see success).
i noted in particular that the lowrisc team has set up as a CIC.
that's a big plus.
technically i am particularly impressed with the concept of using a
32-bit RISCV for GPIO, which they call "minion cores". bitbanging
isn't really bitbanging any more if there's an entire CPU dedicated to
it. the advantage of their approach is that you no longer require
complex multiplexing hardware on the GPIO (as is normally done, with
dedicated hardware blocks for each I/O function). you simply... load
a different program into the minioncore and the pins which e.g. were
previously I2C are now UART. or
some-other-future-as-yet-unspecified-or-unforseen-I/O-interface.
upgrading to the latest version of SDMMC is therefore dead easy: just
write a new program for the minion core.
According to the lowrisc website though, you can make a 64 bit processor
if you so choose though.

http://www.lowrisc.org/about/

I am sure you know this, but I just hope you understand that better
possibilities exist. ;)

I Wonder when they are going to crowdfund though...
Post by Luke Kenneth Casson Leighton
l.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments t
Luke Kenneth Casson Leighton
2017-06-19 01:39:47 UTC
Permalink
Raw Message
Post by zap
According to the lowrisc website though, you can make a 64 bit processor
if you so choose though.
yes. the minioncores would be on the same bus, not requiring any
kind of cache coherency with either each other or with the main 64-bit
CPU(s), either doing DMA writes (on their own) or responding to bus
memory reads/writes in order appear as memory-addressable peripherals.
Post by zap
I am sure you know this, but I just hope you understand that better
possibilities exist. ;)
you may be misunderstanding that the purpose of 32-bit minion cores
is *in addition* to there being one or more main processor(s) which
are SMP or NUMA, which themselves have a bus width (32, 64, 128 bit)
that has absolutely nothing whatsoever to do with the minion-cores
being 32-bit.
Post by zap
I Wonder when they are going to crowdfund though...
i get the general impression that they're quite happy to focus on
development as opposed to raising funds to create an actual processor.
each task requires different skills, time and effort. if the team
focussed on crowdfunding that would be a serious distraction from
their development efforts.

this therefore is an opportunity to create a crowd-funded processor
which utilises their expertise. the one main thing which i could
really do with is a DDR or other high-speed memory interface that is
proven... and compatible with the GPL.

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Se
zap
2017-06-19 02:17:24 UTC
Permalink
Raw Message
Post by Luke Kenneth Casson Leighton
Post by zap
According to the lowrisc website though, you can make a 64 bit processor
if you so choose though.
yes. the minioncores would be on the same bus, not requiring any
kind of cache coherency with either each other or with the main 64-bit
CPU(s), either doing DMA writes (on their own) or responding to bus
memory reads/writes in order appear as memory-addressable peripherals.
Post by zap
I am sure you know this, but I just hope you understand that better
possibilities exist. ;)
you may be misunderstanding that the purpose of 32-bit minion cores
is *in addition* to there being one or more main processor(s) which
are SMP or NUMA, which themselves have a bus width (32, 64, 128 bit)
that has absolutely nothing whatsoever to do with the minion-cores
being 32-bit.
Post by zap
I Wonder when they are going to crowdfund though...
i get the general impression that they're quite happy to focus on
development as opposed to raising funds to create an actual processor.
each task requires different skills, time and effort. if the team
focussed on crowdfunding that would be a serious distraction from
their development efforts.
this therefore is an opportunity to create a crowd-funded processor
which utilises their expertise. the one main thing which i could
really do with is a DDR or other high-speed memory interface that is
proven... and compatible with the GPL.
I guess I misunderstood then on both counts. well this is odd. xD
Post by Luke Kenneth Casson Leighton
l.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@files.phcomp.co.u
J.B. Nicholson
2017-06-19 03:58:53 UTC
Permalink
Raw Message
Post by Luke Kenneth Casson Leighton
indeed. it may sound strange but when there is no other option (and
by that i mean *exhaustive* analysis finds no other option) i do not
mind "crossing the line" into what would traditionally be viewed by
software libre purists as "unacceptable territory" *IF* in doing so it
is part of a long-term strategy to *REPLACE* the very thing being
leveraged [to make money etc. etc.]
Richard Stallman (whose objection to non-free software should require no
explanation) concurs, he has long said that installing and running non-free
software is okay for the purpose of making a free replacement (see
https://www.gnu.org/philosophy/is-ever-good-use-nonfree-program.html for
his essay on this which includes a description of the limits of this
methodology and rationale). Relatedly, he also has told University
audiences that they should have reverse engineering courses (see
https://audio-video.gnu.org/ for recordings of his many speeches). I agree
with him on both points for the same reasons.

After all, once the reverse engineering is complete the free replacement
should suffice to do the practical jobs and then the non-free,
user-subjugating components should be removed in favor of using the free
replacement.
Post by Luke Kenneth Casson Leighton
for example: many software libre supporters flatly refuse to even
*install* Windows NT... but if i had taken that attitude i would not
have broken the NT Domains protocol, over 20 years ago.
I imagine this was and is also true for the Samba team.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachmen
David Niklas
2017-06-23 13:27:31 UTC
Permalink
Raw Message
On Sun Jun 18 07:09:40 BST 2017,
Post by Luke Kenneth Casson Leighton
Post by zap
IF you can figure out how to reverse engineer the 3d engine, ikcl
would be very happy I am sure.
that used to be the case but is no longer true. please read section
https://www.titanians.org/the-bill-of-ethics/
ARM's illegal and unethical activity which destroyed luc verhaegen's
career and reputation - including several counts of slander as well as
blackmail of the company funding his reverse-engineering efforts - was
the last straw.
Well I am kinda stuck with two Mali-GPUs, both I bought under the
impression that the code is open-source if in FLOSS and I would imagine
that there are others in this situation.
Post by Luke Kenneth Casson Leighton
if we proceed to reverse-engineer MALI, logically it results in
people buying more ARM products.
Actually, I read on wikipedia what Linux's support for Mali is and what
the firefly and CHIP webpages described their product as (e.g.
proprietary vs. opensource), and thus I seem to have been fooled, twice.
Surly others will come to the same conclusions unless they dig into this
ML, or the Linux kernel archives, or stumble on Luc's page?
Post by Luke Kenneth Casson Leighton
if people buy more ARM products, logically it results in more money
(resources) going to support ARM's illegal and unethical actions.
Which will happen anyways due to ignorance. I try to stay up on what
companies are nice and which are not plus the general goings on in the
opensource community (I know about Nvidia and AMD/Radeon), and this entire
problem with Mali blindsided me.
This is not a problem that is as well publicized as say, Linux giving
Nvidia the F word.
Post by Luke Kenneth Casson Leighton
any action which is taken that results in support or endorsement of
unethical actions is, logically, itself, unethical
therefore, logically and plainly put: unless ARM's attitude changes
the reverse-engineering of MALI is itself an unethical act.
so it's a simple chain of logical reasoning based on ethical
principles.
l.
Unless you're trying to be merciful to those who are less fortunate.
Look at those who buy prisoners from ISIS or any other violent group.
Or, how about ransomware?
Do we just tell people "Next time don't use windowz"?
It's a tough choice, I agree Luke, but if I succeed you can be certain
that I, like Luc, will pay dearly for my good intentions. Then at least
Luc will gain a friend as I dwell in a state of solidarity with him, and
his actions will not go to waste.

The reason I would really be interested in the Mali GPU driver is not
because of the need for a basic driver, nor for the sake of games
(though they do have an appeal), but because of the Opencl support which
I'd really like to play with (My desktop card is an AMD and their Opencl
support is non-functional at this time in my system for whatever reason
in spite of my using the latest kernel and Mesa library).

Now can we get back to how to do the reverse engineering itself?

Sincerely,
David

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to ar
Bill Kontos
2017-06-23 15:29:32 UTC
Permalink
Raw Message
Post by David Niklas
Now can we get back to how to do the reverse engineering itself?
Someone should suggest Luc to open a monthly patreon-styled donation
page for working on lima.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send
zap
2017-06-23 18:36:48 UTC
Permalink
Raw Message
Post by Bill Kontos
Post by David Niklas
Now can we get back to how to do the reverse engineering itself?
Someone should suggest Luc to open a monthly patreon-styled donation
page for working on lima.
I second this notion. reverse engineering lima could open a lot of
options to us if successful.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to a
Mike Leimon
2017-06-23 19:48:39 UTC
Permalink
Raw Message
Post by Bill Kontos
Post by David Niklas
Now can we get back to how to do the reverse engineering itself?
Someone should suggest Luc to open a monthly patreon-styled donation
page for working on lima.
If ARM is still as openly hostile as they appear to be towards reverse
engineering of their GPU designs, I don't think that pursuing lima
development is a very good option. I suspect that if development of the
driver resumes, they will probably continue their horrible crusade against
Luc and I don't want to see that happen to anybody.

I think that if we are really trying to play the long game here, then that
money would be better spent working towards the development of an
open GPU design that we can can use in future risc-v SOCs. Perhaps
another option that might be more amenable and rewarding for Luc.
Would be to open a monthly patron-styled donation page not to develop
lima drivers, but instead to work together with people like Jeff Bush
(the main guy developming the Nyuzi GPGPU:
https://github.com/jbush001/NyuziProcessor ). To me (and I haven't
read into it much) it seems like, the focus thus far for the Nyuzi project
is only on developing the GPGPU and not so much looking at it from
the point of view of, how would we write a Linux driver for this device.
I think if we focused some money on Luc (driver developer) and the
Nyuzi (hardware developer) developers, then they might be more inspired
to focus their efforts on producing a quality GPGPU and supporting drivers.

Honestly, I think that working together with people who are openly
developing a GPU design and who would probably be quite welcoming
of assistance from Luc would be a much better situation for him than if
he went back to trying to working on the lima driver. Perhaps a company
like si-five ( https://www.sifive.com/ ) or the risc-v organization might
even consider trying to sponsor a collaboration like this as having a good
open GPU implementation would certainly be a boon to them in the future.

Anyhow, that is my two cents.

-Mike
_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send l
Bill Kontos
2017-06-23 20:03:29 UTC
Permalink
Raw Message
Post by Mike Leimon
Honestly, I think that working together with people who are openly
developing a GPU design and who would probably be quite welcoming
of assistance from Luc would be a much better situation for him than if
he went back to trying to working on the lima driver. Perhaps a company
like si-five ( https://www.sifive.com/ ) or the risc-v organization might
even consider trying to sponsor a collaboration like this as having a good
open GPU implementation would certainly be a boon to them in the future.
Anyhow, that is my two cents.
That is a really good idea and luke had mentioned it multiple times in
the past but keep in mind that gpus are an extreme patent minefield
these days. Heck, intel can't solve some problems like memory
compression on their gpus. But maybe someone should propose this to
sifive or the risc v foundation. Something like a Berkeley sponsored
program would be really interesting.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@fi
Bluey
2017-06-24 08:55:28 UTC
Permalink
Raw Message
Post by Bill Kontos
Post by Mike Leimon
Honestly, I think that working together with people who are openly
developing a GPU design and who would probably be quite welcoming
of assistance from Luc would be a much better situation for him than if
he went back to trying to working on the lima driver. Perhaps a company
like si-five ( https://www.sifive.com/ ) or the risc-v organization might
even consider trying to sponsor a collaboration like this as having a good
open GPU implementation would certainly be a boon to them in the future.
Anyhow, that is my two cents.
That is a really good idea and luke had mentioned it multiple times in
the past but keep in mind that gpus are an extreme patent minefield
these days. Heck, intel can't solve some problems like memory
compression on their gpus. But maybe someone should propose this to
sifive or the risc v foundation. Something like a Berkeley sponsored
program would be really interesting.
Please excuse my ignorance in such matters but would it be possible to
use a RISC-V or FPGA chip as an interim eGPU until such time that a more
specialised chip can be developed and released?

I appreciate that RISC-V/FPGA chips are not likely to be well-suited to the task
of GPU processing but perhaps they would be better than no GPU at all.

Once a specialised libre GPU has been developed, the RISC-V / FPGA chips
could be repurposed as a CPU for other projects/computers/laptops/etc. and,
hence, ensuring that they don't go to waste.

- Bluey
_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send
Luke Kenneth Casson Leighton
2017-06-24 13:17:22 UTC
Permalink
Raw Message
Post by Bluey
Please excuse my ignorance in such matters but would it be possible to
use a RISC-V or FPGA chip as an interim eGPU until such time that a more
specialised chip can be developed and released?
on its own (as-is), no. however with certain very very specific and
in some cases specialised SIMD instructions a reasonable approximation
can be had. these operations are:

* SIMD "and" for a bit-wise zero check (as large as possible)
* inverse-squared function (for 1/x^2) - a very common operation in 3D
* SIMD 12-14 bit accurate divide operation.

this latter turns out to be "good enough" for the majority of 3D
operations, where the accuracy on screens which only *have* 1920
pixels (11 bits being sufficient) division calculations beyond 12-14
bits is completely and utterly redundant... *under certain
circumstances*.

the point being that a divide operation which only requires 12-14 bits
of accuracy may complete in half the time, thus dramatically saving on
CPU cycles.
Post by Bluey
I appreciate that RISC-V/FPGA chips are not likely to be well-suited to the task
of GPU processing but perhaps they would be better than no GPU at all.
not "and be power-efficient at the same time"
Post by Bluey
Once a specialised libre GPU has been developed, the RISC-V / FPGA chips
could be repurposed as a CPU for other projects/computers/laptops/etc. and,
hence, ensuring that they don't go to waste.
this was the reasoning behind ICubeCorp's "UPU" - Unified Processing
Unit - which unfortunately they kept proprietary. i tried to help
them to understand the need to release the full boot initialisation
source code and to comply with the GPL but they did not follow up.

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-netb

Luke Kenneth Casson Leighton
2017-06-23 23:27:29 UTC
Permalink
Raw Message
Post by Mike Leimon
If ARM is still as openly hostile as they appear to be towards reverse
engineering of their GPU designs, I don't think that pursuing lima
development is a very good option. I suspect that if development of the
driver resumes, they will probably continue their horrible crusade against
Luc and I don't want to see that happen to anybody.
well forewarned is forearmed. and it's quite easy, when
reverse-engineering, to find security flaws. for every unethical
action that ARM takes it would be really easy to release another
zero-day exploit with full source code and a CVE report.

pretty soon they'd get the message.
Post by Mike Leimon
I think that if we are really trying to play the long game here, then that
money would be better spent working towards the development of an
open GPU design that we can can use in future risc-v SOCs.
exactly my point.

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@files.p
Loading...