- Mar 23, 2013
- 21
- 0
- 18,510
I'm planning on buying an i5 2500k and GTX650ti and I am interested in what is required for the cpu and CUDA functionality to operate at the same time, used by the same program.
Looking at a certain program for computational chemistry (cc being the main reason behind the buy): http://classic.chem.msu.su/gran/gamess/cuding.html an AMD Phenom II X4 955 cannot utilize all cpu cores and all cuda devices at the same time and the sugesstion is that the Phenom does not have Hyper Threading. I understand nVidia gpus may not be the best fit for AMD processors and vice versa but can it really be the HT ? In my oppinion there should not be a connection
in other words if i buy a i5 cpu without HT can i expect the same problem ? these calculations usually take hours to days to complete (on small molecular systems at least) so I am looking at some instructions/day hungry software.
I just want to know is there something in the chipset of different mother boards or in different gpus and cpus that defines wether they can work together (only in the sense that CUDA is enabled) or not.
Can I mix AMD cpu, AMD mb and nVidia CUDA gpu to use all processing power together ? (it seems a cheeper solution and the phenom above is obviously not the latest AMD cpu)
Was it just a problem in the program ? after all this is a new technology of sorts and still highly in testing
what kind of software runs cpus + cuda devices at the same time at max utilization
Looking at a certain program for computational chemistry (cc being the main reason behind the buy): http://classic.chem.msu.su/gran/gamess/cuding.html an AMD Phenom II X4 955 cannot utilize all cpu cores and all cuda devices at the same time and the sugesstion is that the Phenom does not have Hyper Threading. I understand nVidia gpus may not be the best fit for AMD processors and vice versa but can it really be the HT ? In my oppinion there should not be a connection
in other words if i buy a i5 cpu without HT can i expect the same problem ? these calculations usually take hours to days to complete (on small molecular systems at least) so I am looking at some instructions/day hungry software.
I just want to know is there something in the chipset of different mother boards or in different gpus and cpus that defines wether they can work together (only in the sense that CUDA is enabled) or not.
Can I mix AMD cpu, AMD mb and nVidia CUDA gpu to use all processing power together ? (it seems a cheeper solution and the phenom above is obviously not the latest AMD cpu)
Was it just a problem in the program ? after all this is a new technology of sorts and still highly in testing
what kind of software runs cpus + cuda devices at the same time at max utilization
Facebook
Twitter
Instagram