Discussion:
[Arm-netbook] firefly 3399 all source software disclosed?
r***@Safe-mail.net
2017-05-07 09:53:27 UTC
Permalink
https://www.kickstarter.com/projects/1771382379/firefly-rk3399-six-core-64-bit-high-performance-pl
Has all source code for the computer been disclosed?
Lkcl has says, he can make use of the rk3399 if he has the processor's data sheet. https://www.kickstarter.com/projects/1771382379/firefly-rk3399-six-core-64-bit-high-performance-pl/posts/1818477 What data about the rk3399 does lkcl not have?
What is the performance of the rk3399 processor compared to a x86 dual core 2ghz?

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@files.ph
Bill Kontos
2017-05-07 11:13:29 UTC
Permalink
The A72 is supposed to be as fast as a Core M Broadwell core at 2.5 ghz(
assuming that the A72 maintains clockspeed while the Core M throttles,
which is the case in most fanless laptops). The rk3399 is clocked at 2.0
ghz on the A72 cluster so it's slower but it has the benefit of the neon
SIMD extension set for future codecs( should someone want to make it work)
and dedicated ip blocks for a lot of stuff.
Post by r***@Safe-mail.net
https://www.kickstarter.com/projects/1771382379/firefly-
rk3399-six-core-64-bit-high-performance-pl
Has all source code for the computer been disclosed?
Lkcl has says, he can make use of the rk3399 if he has the processor's
data sheet. https://www.kickstarter.com/projects/1771382379/firefly-
rk3399-six-core-64-bit-high-performance-pl/posts/1818477 What data about
the rk3399 does lkcl not have?
What is the performance of the rk3399 processor compared to a x86 dual core 2ghz?
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Luke Kenneth Casson Leighton
2017-05-07 13:06:18 UTC
Permalink
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
Post by r***@Safe-mail.net
https://www.kickstarter.com/projects/1771382379/firefly-rk3399-six-core-64-bit-high-performance-pl
Has all source code for the computer been disclosed?
as far as i was able to ascertain when i last checked: yes. u-boot
and linux kernel, no proprietary pieces at all.
Post by r***@Safe-mail.net
Lkcl has says, he can make use of the rk3399 if he has the processor's data sheet.
i already have the datasheet. it's the reference design that i need.
Post by r***@Safe-mail.net
What data about the rk3399 does lkcl not have?
i need a reference design. full schematics, full pcb layout.
Post by r***@Safe-mail.net
What is the performance of the rk3399 processor compared to a x86 dual core 2ghz?
the rk3288 already outperforms a dual-core intel atom.

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to ar
r***@Safe-mail.net
2017-05-07 20:29:53 UTC
Permalink
All software for the mali-t860 is open source?

-------- Original Message --------
From: Luke Kenneth Casson Leighton <***@lkcl.net>
Apparently from: arm-netbook-***@lists.phcomp.co.uk
To: Linux on small ARM machines <arm-***@lists.phcomp.co.uk>
Subject: Re: [Arm-netbook] firefly 3399 all source software disclosed?
Date: Sun, 7 May 2017 14:06:18 +0100
Post by Luke Kenneth Casson Leighton
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
Post by r***@Safe-mail.net
https://www.kickstarter.com/projects/1771382379/firefly-rk3399-six-core-64-bit-high-performance-pl
Has all source code for the computer been disclosed?
as far as i was able to ascertain when i last checked: yes. u-boot
and linux kernel, no proprietary pieces at all.
Post by r***@Safe-mail.net
Lkcl has says, he can make use of the rk3399 if he has the processor's data sheet.
i already have the datasheet. it's the reference design that i need.
Post by r***@Safe-mail.net
What data about the rk3399 does lkcl not have?
i need a reference design. full schematics, full pcb layout.
Post by r***@Safe-mail.net
What is the performance of the rk3399 processor compared to a x86 dual core 2ghz?
the rk3288 already outperforms a dual-core intel atom.
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-***@file
zap
2017-05-07 21:10:04 UTC
Permalink
Post by r***@Safe-mail.net
All software for the mali-t860 is open source?
probably a good deal of it is, but the question becomes... how much of
it is...
Post by r***@Safe-mail.net
-------- Original Message --------
Subject: Re: [Arm-netbook] firefly 3399 all source software disclosed?
Date: Sun, 7 May 2017 14:06:18 +0100
Post by Luke Kenneth Casson Leighton
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
Post by r***@Safe-mail.net
https://www.kickstarter.com/projects/1771382379/firefly-rk3399-six-core-64-bit-high-performance-pl
Has all source code for the computer been disclosed?
as far as i was able to ascertain when i last checked: yes. u-boot
and linux kernel, no proprietary pieces at all.
Post by r***@Safe-mail.net
Lkcl has says, he can make use of the rk3399 if he has the processor's data sheet.
i already have the datasheet. it's the reference design that i need.
Post by r***@Safe-mail.net
What data about the rk3399 does lkcl not have?
i need a reference design. full schematics, full pcb layout.
Post by r***@Safe-mail.net
What is the performance of the rk3399 processor compared to a x86 dual core 2ghz?
the rk3288 already outperforms a dual-core intel atom.
l.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
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
Luke Kenneth Casson Leighton
2017-05-08 04:45:36 UTC
Permalink
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
Post by r***@Safe-mail.net
All software for the mali-t860 is open source?
none. MALI is proprietary.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send la
Bill Kontos
2017-05-08 09:02:12 UTC
Permalink
I think arm has open source kernel drivers but there is no way they will
get mainlined any time soon. The question is, how much does the userland
blob do and how much work needs to be done to get libre 2d accel.
Post by Luke Kenneth Casson Leighton
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
Post by r***@Safe-mail.net
All software for the mali-t860 is open source?
none. MALI is proprietary.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Luke Kenneth Casson Leighton
2017-05-08 09:11:24 UTC
Permalink
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
I think arm has open source kernel drivers but there is no way they will get
mainlined any time soon. The question is, how much does the userland blob do
everything. the open source kernel "driver" cannot actually be
called a "driver" at all. it is nothing more than a gateway to the
actual hardware. peripherals on ARM are memory-mapped: the "driver"
simply provides access to the shared memory region (plus maybe handles
some interrupts).

this is a constant source of confusion with people believing that
just because the "shim" has been released, somehow magically it's all
okay.

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
S
Bill Kontos
2017-05-08 09:20:56 UTC
Permalink
So do these drivers provide registers etc ? How much extra effort is
required to get 2d accel with a clean room userland driver ? If we can do
that and also get rasterisers working we can have a fully functional
gnu/linux desktop futureproofed with wayland support. Also could we reuse
that multimedia decoding engine from the a20 ?
Post by Luke Kenneth Casson Leighton
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
Post by Bill Kontos
I think arm has open source kernel drivers but there is no way they will
get
Post by Bill Kontos
mainlined any time soon. The question is, how much does the userland
blob do
everything. the open source kernel "driver" cannot actually be
called a "driver" at all. it is nothing more than a gateway to the
actual hardware. peripherals on ARM are memory-mapped: the "driver"
simply provides access to the shared memory region (plus maybe handles
some interrupts).
this is a constant source of confusion with people believing that
just because the "shim" has been released, somehow magically it's all
okay.
l.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Bill Kontos
2017-05-08 09:18:13 UTC
Permalink
Yes, here they are

https://developer.arm.com/products/software/mali-drivers/midgard-kernel
Post by Bill Kontos
I think arm has open source kernel drivers but there is no way they will
get mainlined any time soon. The question is, how much does the userland
blob do and how much work needs to be done to get libre 2d accel.
Post by Luke Kenneth Casson Leighton
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
Post by r***@Safe-mail.net
All software for the mali-t860 is open source?
none. MALI is proprietary.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
m***@gmail.com
2017-05-08 12:54:08 UTC
Permalink
Post by Bill Kontos
I think arm has open source kernel drivers but there is no way they will
get mainlined any time soon. The question is, how much does the userland
blob do and how much work needs to be done to get libre 2d accel.
A display pipeline is a complex one which requires multiple drivers.
Especially on ARM devices where the display pipeline is a "mix and match"
one.

So ARM has released a "driver". What type of driver? What functions does it
have?

Is it one for their display encoder, a crtc driver? A framebuffer driver? A
clasic DRM driver, A KMS driver? A 2D Accelerator for X? A 3D accelator. A
shim to acces the hardware using a BLOB build for one specific version of X
relying and kernel on a specific ARM Core design? Probably the last.

And ARM specific driver without sources is worthless. It's use once and
only for short time.

SoC with ARM desing are mix and match. Different ARM core designs.
Different and mixed GPU's MALI, Vivante, PowerVR, etc. Output (CRTC etc)
from all type of different vendors ARM, Designware, etc. Most of the time
the manufacture is not disclosed.

If you look at Allwinner. The community now has two display engine drivers.
The drivers only setup output and change video modes. GPU type is MALI.

Bealebone has TI SoC for which a KMS/DRM driver was written by Rob clark,
tilcdc, which has support for a NXP HDMI controller/encoder. It also has a
2d accelerator from vivante and a 3d accelatror from Imation(powerVR).

With the recent release of the etnaviv driver the BB community wrote a
patch to have 2d acceleration.

Etnaviv and Freedreno had a hard time being included into the Linux
projects because of their mix and match nature. A single driver was
impossible.

I've once had bought an Atom mini laptop (Dell Mini 1010). Which I
explicitly bought because of the most powerful Intel ATOM GPU at the time.
Intel has a reasonable Opensource GPU track record. Poulsbo turned out to
be an Imation PowerVR licenced design. Boy was I misled: no opensource
drivers only blob's compatible with aging and broken kernels and X-servers.
So no upgrades for me or anyone else.

So a perfectly capable laptop to the trash.

Intel did release a Opensource driver. A KMS driver. So that means
modesetting and nothing more. No 2d acceleration no 3d acceleration nog
hardware video decoding. Intel could not build anything better because
Imation did not let them. The Intel engineer said that there was enough
code and documentation available to create decoder though. But he didn't
write it.

This is way closed source drivers a bad idea. They limit the use and
lifetime of your devices.

I guess the company's building those GPU are too scared that if they open
their drives their competitors see the copyright and/or patent infringement
their design has.

And ARM and Imation are scared the most.

Thankfully Imation is going to die soon. Apple has announced to create
their own GPU. and rob Imation from 60% of their income. Imation anounced a
patent war agains Apple. SCO Unix anyone?
Post by Bill Kontos
Post by Luke Kenneth Casson Leighton
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
Post by r***@Safe-mail.net
All software for the mali-t860 is open source?
none. MALI is proprietary.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Luke Kenneth Casson Leighton
2017-05-08 13:04:45 UTC
Permalink
Post by m***@gmail.com
I guess the company's building those GPU are too scared that if they open
their drives their competitors see the copyright and/or patent infringement
their design has.
eexactly. the "big boys" - nvidia, intel, amd - are the ones less
adversely affected, because their patent portfolio is larger and more
well-established i.e. has precedence.

the "upstarts" are the embedded GPU designers, who, yeah, exactly as
you say, will have almost certainly been infringing various patents.
Post by m***@gmail.com
Thankfully Imation is going to die soon. Apple has announced to create their
own GPU. and rob Imation from 60% of their income. Imation anounced a patent
war agains Apple.
... except somebody forgot to tell their CEO that the patents they
were expecting to bludgeon apple with... have expired. someone at
apple was paying attention, whilst imgtec wasn't.

i want to laugh quite a lot... but it would be unkind to do so, and,
additionally, this is far from over.
Post by m***@gmail.com
SCO Unix anyone?
looking forward to it. it's about time. we _have_ told them again and again

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
S
Bill Kontos
2017-05-08 14:39:17 UTC
Permalink
Imagination is far from dead yet. They did a mistake when they based all of
their gpu income on apple, but their ip is rather advanced. Also on the
plus side there has been discussion about them open sourcing their drivers
a few months back and afterwards they hired new developers. I wouldn't be
surprised if they actually saw this coming and reacted before it happened.
Post by m***@gmail.com
Post by Bill Kontos
I think arm has open source kernel drivers but there is no way they will
get mainlined any time soon. The question is, how much does the userland
blob do and how much work needs to be done to get libre 2d accel.
A display pipeline is a complex one which requires multiple drivers.
Especially on ARM devices where the display pipeline is a "mix and match"
one.
So ARM has released a "driver". What type of driver? What functions does
it have?
Is it one for their display encoder, a crtc driver? A framebuffer driver?
A clasic DRM driver, A KMS driver? A 2D Accelerator for X? A 3D accelator.
A shim to acces the hardware using a BLOB build for one specific version of
X relying and kernel on a specific ARM Core design? Probably the last.
And ARM specific driver without sources is worthless. It's use once and
only for short time.
SoC with ARM desing are mix and match. Different ARM core designs.
Different and mixed GPU's MALI, Vivante, PowerVR, etc. Output (CRTC etc)
from all type of different vendors ARM, Designware, etc. Most of the time
the manufacture is not disclosed.
If you look at Allwinner. The community now has two display engine
drivers. The drivers only setup output and change video modes. GPU type is
MALI.
Bealebone has TI SoC for which a KMS/DRM driver was written by Rob clark,
tilcdc, which has support for a NXP HDMI controller/encoder. It also has a
2d accelerator from vivante and a 3d accelatror from Imation(powerVR).
With the recent release of the etnaviv driver the BB community wrote a
patch to have 2d acceleration.
Etnaviv and Freedreno had a hard time being included into the Linux
projects because of their mix and match nature. A single driver was
impossible.
I've once had bought an Atom mini laptop (Dell Mini 1010). Which I
explicitly bought because of the most powerful Intel ATOM GPU at the time.
Intel has a reasonable Opensource GPU track record. Poulsbo turned out to
be an Imation PowerVR licenced design. Boy was I misled: no opensource
drivers only blob's compatible with aging and broken kernels and X-servers.
So no upgrades for me or anyone else.
So a perfectly capable laptop to the trash.
Intel did release a Opensource driver. A KMS driver. So that means
modesetting and nothing more. No 2d acceleration no 3d acceleration nog
hardware video decoding. Intel could not build anything better because
Imation did not let them. The Intel engineer said that there was enough
code and documentation available to create decoder though. But he didn't
write it.
This is way closed source drivers a bad idea. They limit the use and
lifetime of your devices.
I guess the company's building those GPU are too scared that if they open
their drives their competitors see the copyright and/or patent infringement
their design has.
And ARM and Imation are scared the most.
Thankfully Imation is going to die soon. Apple has announced to create
their own GPU. and rob Imation from 60% of their income. Imation anounced a
patent war agains Apple. SCO Unix anyone?
Post by Bill Kontos
Post by Luke Kenneth Casson Leighton
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
Post by r***@Safe-mail.net
All software for the mali-t860 is open source?
none. MALI is proprietary.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Lauri Kasanen
2017-05-08 14:52:45 UTC
Permalink
On Mon, 8 May 2017 17:39:17 +0300
Post by Bill Kontos
Imagination is far from dead yet. They did a mistake when they based all of
their gpu income on apple, but their ip is rather advanced. Also on the
plus side there has been discussion about them open sourcing their drivers
a few months back and afterwards they hired new developers. I wouldn't be
surprised if they actually saw this coming and reacted before it happened.
They're openly considering selling MIPS, as well as other parts of
them. That goes against knowing things would go this way.

- Lauri

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large attachments to arm-***@files.
m***@gmail.com
2017-05-08 15:26:17 UTC
Permalink
2017-05-08 16:39 GMT+02:00 Bill Kontos <***@gmail.com>:

Please don't top post. Gmail has inline capability.
Post by Bill Kontos
Imagination is far from dead yet. They did a mistake when they based all
of their gpu income on apple, but their ip is rather advanced.
ImgTec income is going to be cut by 60%. That's not something most
companies can survive.

Their plan is to sue Apple for using of their IP and forcing them te keep
paying royalties.

To fund that they are selling of vital parts, like the recently aquired
MIPS.
, to fund the legal fees and supplement their loss of income.

Apple will stall, they have the money to do so, and probably force them to
trial. In trail ImgTec is forced to disclose publicly what they think
belongs to them. Which they will be reluctant to do because then the other
GPU manufactures might be able file suits for infringement to ImgTec.

Hence my reference to the SCO suit against Linux users. Which ultimately
failed and killed the company.

The same issue for them is they open up their driver sources.

All GPU manufactures are bound to infringement because sometimes things are
invented twice or more at different places. Which makes sense when you're
trying to reach the same goal with the same means and knowledge.

If only they'd sit around and say let's all open up our drivers and not sue
each other for current design and start over from there.

I'll be surprised if they'll survive this. Apple has probably tried to buy
and failed so they move to kill. This is join us or die tactics. And Apple
has more money thus is able to strike harder and longer.

Also on the plus side there has been discussion about them open sourcing
Post by Bill Kontos
their drivers a few months back and afterwards they hired new developers. I
wouldn't be surprised if they actually saw this coming and reacted before
it happened.
ImgTec has been holding that bone for our noses for years. Ten to be exact
Poulsbo dates from 2007. I don't believe it until it's proven.

So if I'd place a bet to
1. ImgTec releasing sources
2. ImgTec dying from Apple lawsuit

I'd place on the second one. Their selling long term part for short term
money for a lost cause.

The only winning parties here are going to be the lawyers just like with
SCO.
r***@Safe-mail.net
2017-05-08 15:27:28 UTC
Permalink
Best knowledge is, that new intel and amd processors cannot be reverse engineered.
What in regard of the latest mali gpus?
If you have the money, they can be reverse engineered?

-------- Original Message --------
From: Lauri Kasanen <***@gmx.com>
Apparently from: arm-netbook-***@lists.phcomp.co.uk
To: Linux on small ARM machines <arm-***@lists.phcomp.co.uk>
Subject: Re: [Arm-netbook] firefly 3399 all source software disclosed?
Date: Mon, 8 May 2017 17:52:45 +0300
Post by Lauri Kasanen
On Mon, 8 May 2017 17:39:17 +0300
Post by Bill Kontos
Imagination is far from dead yet. They did a mistake when they based all of
their gpu income on apple, but their ip is rather advanced. Also on the
plus side there has been discussion about them open sourcing their drivers
a few months back and afterwards they hired new developers. I wouldn't be
surprised if they actually saw this coming and reacted before it happened.
They're openly considering selling MIPS, as well as other parts of
them. That goes against knowing things would go this way.
- Lauri
_______________________________________________
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.c
Luke Kenneth Casson Leighton
2017-05-08 15:43:07 UTC
Permalink
Post by r***@Safe-mail.net
Best knowledge is, that new intel and amd processors cannot be reverse engineered.
What in regard of the latest mali gpus?
If you have the money, they can be reverse engineered?
yes. about $150k would do it. but the question is, really: what
would happen if you did? and, what else could you do with the same
money?

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. any company tries to claim patent
royalties, all that happens is a search is made on their "claims", for
anything similar that has prior art.

if it's another company, guess what? we notify that other company
and watch the fireworks...

l.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send
Allan Mwenda
2017-05-09 05:24:02 UTC
Permalink
Quick give this man $150K for a libre CPU
Post by r***@Safe-mail.net
Post by r***@Safe-mail.net
Best knowledge is, that new intel and amd processors cannot be
reverse engineered.
Post by r***@Safe-mail.net
What in regard of the latest mali gpus?
If you have the money, they can be reverse engineered?
yes. about $150k would do it. but the question is, really: what
would happen if you did? and, what else could you do with the same
money?
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. any company tries to claim patent
royalties, all that happens is a search is made on their "claims", for
anything similar that has prior art.
if it's another company, guess what? we notify that other company
and watch the fireworks...
l.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
Luke Kenneth Casson Leighton
2017-05-09 09:03:48 UTC
Permalink
Post by Allan Mwenda
Quick give this man $150K for a libre CPU
:)

_______________________________________________
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
r***@Safe-mail.net
2017-05-09 16:13:53 UTC
Permalink
-------- Original Message --------
From: Allan Mwenda <***@gmail.com>
Apparently from: arm-netbook-***@lists.phcomp.co.uk
To: Linux on small ARM machines <arm-***@lists.phcomp.co.uk>, Luke Kenneth Casson Leighton <***@lkcl.net>
Subject: Re: [Arm-netbook] firefly 3399 all source software disclosed?
Date: Tue, 09 May 2017 08:24:02 +0300
Post by Allan Mwenda
Quick give this man $150K for a libre CPU
Post by r***@Safe-mail.net
Best knowledge is, that new intel and amd processors cannot be reverse engineered.
What in regard of the latest mali gpus?
If you have the money, they can be reverse engineered?
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
How can you make an arm gpu for 150000usd? How would you use it? Would you put it additionally on a bord and not use the gpu located on the processor socket?
If you reverse engineer the latest mali gpu, what you hold against it is, that when next version of an arm socket is for sale it will have a new mali gpu and require a new reverse engineering?
Lets say 50000 people would buy the filrefly rk3399. Same people would pay 3eu each for reverse engineering the gpu. Then we would have a source code arm computer?
It takes a lot of coordination to provide gnulinux distributions I assume. It is unfortunate people are not able to coordinate such that we can get source code hardware.

_______________________________________________
arm-netbook mailing list arm-***@lists.phcomp.co.uk
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Send large att
Luke Kenneth Casson Leighton
2017-05-09 16:26:58 UTC
Permalink
Post by r***@Safe-mail.net
How can you make an arm gpu for 150000usd?
you don't. ARM is a registered trademark, and copying what they're
doing is asking for trouble.

instead you take one of the "open gpus" or parts of them and use that.
there's several i've been tracking: MIAOW, Nyuzi, the ORSOC graphics
accelerator - there's surprisingly quite a lot out there.
Post by r***@Safe-mail.net
How would you use it?
it would be on the same silicon, on the same memory bus as the RISC-V
64-bit core. anything else is too power-hungry.
Post by r***@Safe-mail.net
Would you put it additionally on a bord and not use the gpu located on the processor socket?
no.
Post by r***@Safe-mail.net
If you reverse engineer the latest mali gpu, what you hold against it is, that when next version of an arm socket is for sale it will have a new mali gpu and require a new reverse engineering?
correct. so whatever you get it's guaranteed to be "old". this is
the sad fact of reverse-engineering: all that effort, with *no
guarantee of success*.... just to get something that's years
out-of-date.
Post by r***@Safe-mail.net
Lets say 50000 people would buy the filrefly rk3399. Same people would pay 3eu each for reverse engineering the gpu. Then we would have a source code arm computer?
no. ARM has acted so unethically in slandering luc verhaegen and
blackmailing companies that pay him that i am not interested in
supporting their business any more than is absolutely necessary.
Post by r***@Safe-mail.net
It takes a lot of coordination to provide gnulinux distributions I assume.
yes. a _lot_. however amazingly people are actually doing just
that... even on risc-v where there isn't yet even any actual hardware.
Post by r***@Safe-mail.net
It is unfortunate people are not able to coordinate such that we can get source code hardware.
it's still quite a young area of focus.

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.phcomp
Allan Mwenda
2017-05-10 07:04:49 UTC
Permalink
Hi Luke
On the topic of GPU's have you seen this?
https://github.com/etnaviv/etna_viv
Seems to be a libre driver for Vivante GPU's
Post by Luke Kenneth Casson Leighton
Post by r***@Safe-mail.net
How can you make an arm gpu for 150000usd?
you don't. ARM is a registered trademark, and copying what they're
doing is asking for trouble.
instead you take one of the "open gpus" or parts of them and use that.
there's several i've been tracking: MIAOW, Nyuzi, the ORSOC graphics
accelerator - there's surprisingly quite a lot out there.
Post by r***@Safe-mail.net
How would you use it?
it would be on the same silicon, on the same memory bus as the RISC-V
64-bit core. anything else is too power-hungry.
Post by r***@Safe-mail.net
Would you put it additionally on a bord and not use the gpu located on
the processor socket?
no.
Post by r***@Safe-mail.net
If you reverse engineer the latest mali gpu, what you hold against it
is, that when next version of an arm socket is for sale it will have a new
mali gpu and require a new reverse engineering?
correct. so whatever you get it's guaranteed to be "old". this is
the sad fact of reverse-engineering: all that effort, with *no
guarantee of success*.... just to get something that's years
out-of-date.
Post by r***@Safe-mail.net
Lets say 50000 people would buy the filrefly rk3399. Same people would
pay 3eu each for reverse engineering the gpu. Then we would have a source
code arm computer?
no. ARM has acted so unethically in slandering luc verhaegen and
blackmailing companies that pay him that i am not interested in
supporting their business any more than is absolutely necessary.
Post by r***@Safe-mail.net
It takes a lot of coordination to provide gnulinux distributions I
assume.
yes. a _lot_. however amazingly people are actually doing just
that... even on risc-v where there isn't yet even any actual hardware.
Post by r***@Safe-mail.net
It is unfortunate people are not able to coordinate such that we can get
source code hardware.
it's still quite a young area of focus.
l.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
Luke Kenneth Casson Leighton
2017-05-10 07:12:57 UTC
Permalink
Post by Allan Mwenda
Hi Luke
On the topic of GPU's have you seen this?
https://github.com/etnaviv/etna_viv
yeahyeah i have.
Post by Allan Mwenda
Seems to be a libre driver for Vivante GPU's
it is. the problem with vivante is that their *own* proprietary
driver is so shit that the entire reputation of vivante's technology
has been shot to shit.

i spoke with one fabless semiconductor company, urging them to use
etnaviv: unfortunately my contact used the word "vivante" when talking
to the engineer, who told my contact "it's shit. it keeps crashing.
so we won't use it".

whoopsie...

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.phcomp.
Allan Mwenda
2017-05-10 08:05:25 UTC
Permalink
Oh that sux
Is the gpu itself fairly capable though?
Post by Luke Kenneth Casson Leighton
Post by Allan Mwenda
Hi Luke
On the topic of GPU's have you seen this?
https://github.com/etnaviv/etna_viv
yeahyeah i have.
Post by Allan Mwenda
Seems to be a libre driver for Vivante GPU's
it is. the problem with vivante is that their *own* proprietary
driver is so shit that the entire reputation of vivante's technology
has been shot to shit.
i spoke with one fabless semiconductor company, urging them to use
etnaviv: unfortunately my contact used the word "vivante" when talking
to the engineer, who told my contact "it's shit. it keeps crashing.
so we won't use it".
whoopsie...
l.
_______________________________________________
http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
Continue reading on narkive:
Search results for '[Arm-netbook] firefly 3399 all source software disclosed?' (Questions and Answers)
4
replies
Can someone help me out with these words?
started 2008-09-14 07:08:52 UTC
words & wordplay
Loading...