AMD Catalyst 10.9 to Release 15/9/10-usa time

[jrch2k8]

"2.) opencl, well it seems decent but some of my devs tested it regularly and nvidia still wipe the floor with fglrx in many use scenerios, so for now quadro are here to stay in my particular case"

you know hd4xxx emulate 1 opencl gpu cache in vram so a hd4xxx card can't beat a nvidia ever.
i hope you do your testing with a hd5000 card.

and wine is still nvidia focused but i think on dx10 games this will be better because on dx9 there are many renderpaths and on dx10 there are only1 renderpath..

all test are made with a firepro V7800 2GB, im pretty sure this card is based on 5000 series but is true that im not sure if this bug prevail in this card. on nvidia we use a Quadro FX 3700(is not a tesla system but is good enough and accesible for my clients)

about wine dx10 is almost(few less glitches) the same but slower and i have corruption on native games too, which i dont have with radeon or nvidia or nouveau. and well many ppl are freaking out with hon support too(i dont have it, maybe someone will post how it went with 10.9)

[jrch2k8]

as a note im not telling that nvidia is faster than amd at hardware level, im just saying that for my specific needs nvidia behave much better and stable than amd at software level and since some of my apps are critical i just cannot afford to use something like fglrx(even if nvidia have many issues too is very stable in controlled enviroments but i can't just make that happen with fglrx and redhat/novell products for my specific needs <not a fan of those distros but enterprise users want enterprise distros >).

im not implying either that fglrx opencl is bad, to me it could be that opencl is working over cuda on nvidia wich is a tech more mature and tested for the big sharks and well amd opencl/stream is very new(at least i've never seen a product based on stream on the consumer or enterprise market, i can be wrong ofc)

[Qaridarium]

all test are made with a firepro V7800 2GB, im pretty sure this card is based on 5000 series but is true that im not sure if this bug prevail in this card. on nvidia we use a Quadro FX 3700(is not a tesla system but is good enough and accesible for my clients)

a 5000 series card is ok to compare.. but remember openCL is build by nvidia+apple no big clue to be faster for nvidia ;-)

about wine dx10 is almost(few less glitches) the same but slower and i have corruption on native games too, which i dont have with radeon or nvidia or nouveau. and well many ppl are freaking out with hon support too(i dont have it, maybe someone will post how it went with 10.9)

sorry my dx10 argument was not an status report just an remember for the dx9 renderpath mess vs an clear renderpath in dx10.

[Qaridarium]

as a note im not telling that nvidia is faster than amd at hardware level, im just saying that for my specific needs nvidia behave much better and stable than amd at software level and since some of my apps are critical i just cannot afford to use something like fglrx(even if nvidia have many issues too is very stable in controlled enviroments but i can't just make that happen with fglrx and redhat/novell products for my specific needs <not a fan of those distros but enterprise users want enterprise distros >).

im not implying either that fglrx opencl is bad, to me it could be that opencl is working over cuda on nvidia wich is a tech more mature and tested for the big sharks and well amd opencl/stream is very new(at least i've never seen a product based on stream on the consumer or enterprise market, i can be wrong ofc)

i'm thankful to nvidia+apple to build something like openCL and if you use openCL for your work also on an nvidia that is ok because you get an real chance to run the same code on an amd card in the future.

stream+cuda are just an old storry about an cold war in the past.


intel for exampel do not wana openCL because they wana make more money on cpus..

[Epsilon]

i'm thankful to nvidia+apple to build something like openCL and if you use openCL for your work also on an nvidia that is ok because you get an real chance to run the same code on an amd card in the future.

stream+cuda are just an old storry about an cold war in the past.


intel for exampel do not wana openCL because they wana make more money on cpus..

What are you talking about? get your facts straight.

OpenCL was initially developed by Apple Inc., which holds trademark rights, and refined into an initial proposal in collaboration with technical teams at AMD, IBM, Intel, and Nvidia. Apple submitted this initial proposal to the Khronos Group. On June 16, 2008 the Khronos Compute Working Group was formed[1] with representatives from CPU, GPU, embedded-processor, and software companies. This group worked for five months to finish the technical details of the specification for OpenCL 1.0 by November 18, 2008.[2] This technical specification was reviewed by the Khronos members and approved for public release on December 8, 2008.

[3]

Members of Khronos Group

In 2006, some of the OpenGL ARB Working Group members were:
AMD/ATI (AMD acquired ATI)
Apple Inc.
ARM Holdings
Creative Labs
id Software
Ericsson
Google
Intel Corporation
Motorola
Mozilla
Nokia
Nvidia
Samsung Electronics
Sony Computer Entertainment
Oracle/Sun Microsystems (Oracle acquired Sun Microsystems)
Texas Instruments

[Qaridarium]

What are you talking about? get your facts straight.

[3]

Members of Khronos Group

" Khronos Group" is not openCL intel is in the "Khronos Group" because of openGL.

show me the code of using openCL on an intel GPU!

[Epsilon]

Learn to read or stop being a troll.

[Qaridarium]

Learn to read or stop being a troll.

just show me the code openCL on intel GPU please.

or OpenCL on intel CPU without amd tools.