Which forum CUDA programs are DP vs SP?

VBCurtis · 2017-01-09, 07:08

I am considering the purchase of a secondhand workstation with a Quadro 5000 GPU, which has DP performance = 1/2 SP performance. If I purchase this machine, I surely do not want to waste this GPU running single-precision calculations among my choices at mersenneforum.

Which of our CUDA-enabled programs require DP calculations? Which are SP? I already use msieve poly select, ECM, cudaLLR; I would consider trial-factoring, cudalucas, and perhaps something I'm not thinking of (a sieve for a project I already like, say).

Mark Rose · 2017-01-09, 07:47

Keep in mind the Quadro 5000 is about as fast as a GTX 960 for LL. It is still reasonably power efficient when doing DP though.

preda · 2017-04-13, 07:29

TF (mfaktc, mfakto) is SP.
LL (CUDALucas, clLucas) is DP.

LL is also likely limited by memory bandwidth, not only DP performance.

kriesel · 2017-12-26, 18:50

Quote:

Originally Posted by preda

TF (mfaktc, mfakto) is SP.
LL (CUDALucas, clLucas) is DP.

LL is also likely limited by memory bandwidth, not only DP performance.

P-1 (CUDAPM1) is also DP. A brief look at the v0.20 source shows lots of variables typed double. Floats are rare by comparison. GPU Memory occupancy I've observed via GPU-Z is P-1>LL>TF. Occupancy limits maximum exponent for a given GPU's memory size, while bandwidth affects execution speed. But it seems reasonable to expect bandwidth and occupancy to be related. Even a 1GB equipped GPU (Quadro2000 for example) can run exponents well above the current Primenet wavefronts.

2017-01-09, 07:08	#1
VBCurtis "Curtis" Feb 2005 Riverside, CA 2×2,927 Posts	Which forum CUDA programs are DP vs SP? I am considering the purchase of a secondhand workstation with a Quadro 5000 GPU, which has DP performance = 1/2 SP performance. If I purchase this machine, I surely do not want to waste this GPU running single-precision calculations among my choices at mersenneforum. Which of our CUDA-enabled programs require DP calculations? Which are SP? I already use msieve poly select, ECM, cudaLLR; I would consider trial-factoring, cudalucas, and perhaps something I'm not thinking of (a sieve for a project I already like, say).

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Seg fault with various programs	Romuald	Factoring	3	2016-11-14 05:38
Sieving programs	calimero22	Software	12	2015-11-22 08:43
they used to be called programs...	chappy	Lounge	15	2012-08-11 21:02
Compiling CUDA programs on VS 2010 Express	ET_	GPU Computing	9	2011-03-24 10:11
Two programs on same machine?	Unregistered	Software	14	2004-02-15 16:36

2017-01-09, 07:47	#2
Mark Rose "/X\(‘-‘)/X\" Jan 2013 https://pedan.tech/ 6160₈ Posts	Keep in mind the Quadro 5000 is about as fast as a GTX 960 for LL. It is still reasonably power efficient when doing DP though.

2017-04-13, 07:29	#3
preda "Mihai Preda" Apr 2015 2²×3×11² Posts	TF (mfaktc, mfakto) is SP. LL (CUDALucas, clLucas) is DP. LL is also likely limited by memory bandwidth, not only DP performance.