[BOUNTY] sha256 shader for Linux OSS video drivers (15 BTC pledged)

[jgarzik]

Become an open source hero, and help bitcoin too!

OK, I think this project would see some real return (in BTC) on Linux, for all the miners out there.  It would benefit open source as well.

The Project
-------------------------------------------------------
Successfully load and execute a sha256 "compute shader", using 100% open source video drivers on Linux (using closed source ATI tools to produce shader binary is permitted).  Any Linux OS/distribution, as long as it's a recent version.  Must work on ATI 5870/5970 hardware.


Rationale
-------------------------------------------------------
1. In theory, the closed source ATI SDK and video driver should not be needed, once we have a compiled shader.  It would make life much easier on Linux, and expand our miner base, if stock open source drivers can be used for GPU mining.

2. Open source GPGPU efforts are moving slowly, and this would help jump-start those efforts, by providing a working example.  This has the potential to be a high profile contribution to the OSS community.


Details
-------------------------------------------------------
According to some knowledgeable hackers, it should be possible to upload a "compute shader" using current Linux/OSS video drivers, via the Linux DRI APIs.  The programmer (or team) would need to figure out how to coax ATI's SDK to produce a compiled, binary object that is then loaded into an open source driver, and executed.

The person or team collecting this bounty will need to be able to accomplish tasks such as rebuilding and replacing the kernel, rebuilding and replacing Mesa (OpenGL/DRI), and rebuilding/replacing the X server.  Even though these are non-programming tasks, they are decidedly non-trivial.

This code (from ATI?) should be helpful in demonstrating how to work with 5870/5970 hardware: http://cgit.freedesktop.org/mesa/r600_demo/tree/?h=master

Although this task should be largely a "put together existing pieces and make them work" task, it is still quite complex.


The Pledges (in BTC)
-------------------------------------------------------
I'm hoping to raise at least 200 BTC for this task, if not more.  Miners on Linux, consider pledging a block (or part of a block).

15     jgarzik


If you wish to pledge anonymously, send me a PM and I'll coordinate.

Pledges should be payable within 24 hours of a working example being posted publicly.

[Luke-Jr]

I'm offering 50 BTC to the first only-open-source miner to achieve a minimum of 252 MH/s (that's 95% of my present 265 MH/s) on my Radeon 5850. To claim, please send me an email at luke+openminingbounty@dashjr.org with the SHA256 hash of your miner tbz2, in case this turns out to be a close race.

Edit: This offer is expired.

[jgarzik]

Bump.  Increased bounty to 400 BTC.

Here's a link showing several examples of asm shaders on ATI:
http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/tree/src/evergreen_shader.c

and here is some useful Mesa code for building asm shaders:
http://cgit.freedesktop.org/mesa/mesa/tree/src/mesa/drivers/dri/r600/r700_assembler.c
http://cgit.freedesktop.org/mesa/mesa/tree/src/mesa/drivers/dri/r600/r700_shader.c

[Mahkul]

I will pledge 25 BTC for this.

[xf2_org]

Bump, for the new arrivals. 

[Zamicol]

Its not much, but I'll pledge 10 BTC.  Evey little bit(coin) counts right? 

[teknohog]

Here's another 10 BTC. With the recent USD price of bitcoins, I wouldn't even say "it's not much".

[xf2_org]

bump, and updated first post to include the two most recent pledges (from past 30 days).

At today's exchange rate, the bounty is over $3,100.

[DiabloD3]

Become an open source hero, and help bitcoin too!

OK, I think this project would see some real return (in BTC) on Linux, for all the miners out there.  It would benefit open source as well.

The Project
-------------------------------------------------------
Successfully load and execute a sha256 "compute shader", using 100% open source video drivers on Linux (using closed source ATI tools to produce shader binary is permitted).  Any Linux OS/distribution, as long as it's a recent version.  Must work on r600/r700 ATI hardware.

This is pointless. R600 cannot run compute shaders of the kind we need, R700 (Radeon 4xxx) suck at it.

Also, Mesa has a prototype OpenCL compiler for Gallium targets. Your "bounty" is technically already completed before you started.

If you really want to help open source, go work on that project instead.

[MoonShadow]

I'm not a programmer, so I don't know what I'm talking about here, but could such a linux binary permit graphics hardware too old to use the current miners to contribute at a respectable hash/watt rate even if they such at the hash/second rate?

[xf2_org]

This is pointless. R600 cannot run compute shaders of the kind we need, R700 (Radeon 4xxx) suck at it.

That was a reference to the starting point for the architecture in the source code.  You find R700/R800/+ hardware support code in directories labelled "r600" due to several similarities.

Also, Mesa has a prototype OpenCL compiler for Gallium targets. Your "bounty" is technically already completed before you started.

If that was true, then somebody would have collected the $3000+ in free money.

[DiabloD3]

If that was true, then somebody would have collected the $3000+ in free money.

Not at all. Xorg nor FDO accept donations in BTC.

[DiabloD3]

I'm not a programmer, so I don't know what I'm talking about here, but could such a linux binary permit graphics hardware too old to use the current miners to contribute at a respectable hash/watt rate even if they such at the hash/second rate?

No. They lack the hardware design to run programs like this, plus they would be amazingly slow.

[MoonShadow]

I'm not a programmer, so I don't know what I'm talking about here, but could such a linux binary permit graphics hardware too old to use the current miners to contribute at a respectable hash/watt rate even if they such at the hash/second rate?

No. They lack the hardware design to run programs like this, plus they would be amazingly slow.

Slow is irrelevant, if they are efficient.  There are millions of them.  But if it's not possible, it's not possible.

[xf2_org]

Let's not get distracted by Diablo getting worked up over a directory name.

The bounty is for working on 5870/5970/6990 era hardware.

[Basiley]

anyone with [deprecated]Render Monkey experience ?
note: using shader allow you to extract MORE from you GPU. also CAL implementation might me considerably faster than OpenCL one[why, how and some optimisation tricks was nuff said by BarsWF creator already].
and such experience also handy for porting solver in other areas[Linpack/Livermore shader ? Biology ? Meteorology ? Radiophysic]

[DiabloD3]

anyone with [deprecated]Render Monkey experience ?
note: using shader allow you to extract MORE from you GPU. also CAL implementation might me considerably faster than OpenCL one[why, how and some optimisation tricks was nuff said by BarsWF creator already].
and such experience also handy for porting solver in other areas[Linpack/Livermore shader ? Biology ? Meteorology ? Radiophysic]

ArtForz has a CAL miner that is slightly faster than the CL kernel poclbm and I based ours on. Its not particularly a huge win, especially when SDK 2.5 is getting rid of CAL support.

Also, I have experimented with GLSL-based miners. The lack of real integer support kills the whole thing.

[Basiley]

:-(

[DiabloD3]

:-(

Well, its why OpenCL was invented in the first place. Using GLSL for even generic computing tasks that don't fit the OpenGL workflow is problematic and really not worth it.

I don't want to shoot down anyone's hopes, but Mesa already is growing OpenCL support for Gallium. What more could we possibly ask for?

[Basiley]

:-(

Well, its why OpenCL was invented in the first place. Using GLSL for even generic computing tasks that don't fit the OpenGL workflow is problematic and really not worth it.

I don't want to shoot down anyone's hopes, but Mesa already is growing OpenCL support for Gallium. What more could we possibly ask for?

best wishes ?
quicker Gallium3D adoption ?
quicker Mesa development[along with Gallium3D] ?
better drivers[esp free drivers. now about 10x times slower than proprietary counterparts] ?
more suitable SDK' ?
AMD/NVidia support for both developers and OpenCL itself[today Intel CPU's had better OpenCL support than Nvidia GPU's].
and yes OpenCL/OpenGL ES is cool. at least in theory. heil glorious OpenMAXdeveloper