Article: Parallelism at HotPar 2010
By: AM (myname4rwt.delete@this.jee-male.com), August 3, 2010 11:47 pm
Room: Moderated Discussions
anonymous (no@spam.com) on 8/3/10 wrote:
---------------------------
>>Here is a very simple reality check for you (and David): get a machine with win7
>>on and check how fast warp can render, say, Crysis (use the benchmark tool). GTX
>>460 (available from $200 these days) cranks out over 30 fps in 1680x1050, VHD and
>>over 60 fps (GASP) in SLI, same mode. And very short of 30/60 fps in 1920x1080, VHD from the report I saw.
>>
>>How fast do you think CPU can handle this task (btw, a representative of a very
>>widespread class of workloads), even the Intel's 6 core you mentioned? And how many
>>Intel's 6-core crown jewels selling for $1k+ a pop will it take to get the same performance?
>>
>>Have a nice time reevaluating your claims (or better yet, running the test and reporting the results).
>
>I got an even better reality check for you: tell me the SPECint and SPECfp.
>
>I have yet to see GPGPU work well for problems that aren't parallel math.
>
>Like uhm, you know... the bulk of the code that people are dealing with.
You should probably read at least a few posts up the thread before posting. The claims made were
a) reported 100x [and above] speedups are crap (Kanter).
b) *real* performance advantage of GPUs is 2.5x-5x if compared against well-optimized CPU code (Mark Roulo).
While both *may* be true indeed (very easily for codes utilizing a very small portion of GPU's hw), both are infinetely far from being a universal truth.
---------------------------
>>Here is a very simple reality check for you (and David): get a machine with win7
>>on and check how fast warp can render, say, Crysis (use the benchmark tool). GTX
>>460 (available from $200 these days) cranks out over 30 fps in 1680x1050, VHD and
>>over 60 fps (GASP) in SLI, same mode. And very short of 30/60 fps in 1920x1080, VHD from the report I saw.
>>
>>How fast do you think CPU can handle this task (btw, a representative of a very
>>widespread class of workloads), even the Intel's 6 core you mentioned? And how many
>>Intel's 6-core crown jewels selling for $1k+ a pop will it take to get the same performance?
>>
>>Have a nice time reevaluating your claims (or better yet, running the test and reporting the results).
>
>I got an even better reality check for you: tell me the SPECint and SPECfp.
>
>I have yet to see GPGPU work well for problems that aren't parallel math.
>
>Like uhm, you know... the bulk of the code that people are dealing with.
You should probably read at least a few posts up the thread before posting. The claims made were
a) reported 100x [and above] speedups are crap (Kanter).
b) *real* performance advantage of GPUs is 2.5x-5x if compared against well-optimized CPU code (Mark Roulo).
While both *may* be true indeed (very easily for codes utilizing a very small portion of GPU's hw), both are infinetely far from being a universal truth.
