Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
Message boards :
Sieving :
Factorial Sieving discussion
| Author |
Message |
|---|
rogueVolunteer developer
Send message
Joined: 8 Sep 07
Posts: 1256
ID: 12001
Credit: 18,565,548
RAC: 0
 
|
|
I hope to have a factorial sieve for the GPU available (OpenCL) for testing in the next week. It will be at least 10x faster than the CPU sieve. I expect at least 25x faster. | |
|
|
|
I hope to have a factorial sieve for the GPU available (OpenCL) for testing in the next week. It will be at least 10x faster than the CPU sieve. I expect at least 25x faster.
I will gladly test this if it can run on a GTX 570.
____________
 | |
|
|
|
|
4516G-4520G K_Lewis Completed (14 factors, i7 2600K @ 3400+ MHz, Windows OS7 Home 64-bit, -t4, ~49 kp/s)
I hope to have a factorial sieve for the GPU available (OpenCL) for testing in the next week. It will be at least 10x faster than the CPU sieve. I expect at least 25x faster.
I was aware, it's what prompted me to understand this process first of all. I played around with the Mac and Windows builds of what you're creating on a few old GPU's and I'm very much looking forward to its release :)
____________
147*2^1392930+1 was my first prime number found, others have followed :) | |
|
rogueVolunteer developer
Send message
Joined: 8 Sep 07
Posts: 1256
ID: 12001
Credit: 18,565,548
RAC: 0
|
4516G-4520G K_Lewis Completed (14 factors, i7 2600K @ 3400+ MHz, Windows OS7 Home 64-bit, -t4, ~49 kp/s)
I hope to have a factorial sieve for the GPU available (OpenCL) for testing in the next week. It will be at least 10x faster than the CPU sieve. I expect at least 25x faster.
I was aware, it's what prompted me to understand this process first of all. I played around with the Mac and Windows builds of what you're creating on a few old GPU's and I'm very much looking forward to its release :)
I'm running into a problem. The code runs, but the computer becomes unresponsive if I size of work is too large. If I make it smaller, then the speed improvements are lost. The problem is that the computation requires a tight loop in the GPU. I need to break it up into smaller chunks of work. I have some ideas. Hopefully I can work on it tomorrow. | |
|
 Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 435,681,308
RAC: 871,866
                              
|
|
Moving a few posts here from the reservation thread...
____________
My lucky number is 75898524288+1 | |
|
HAmstyVolunteer tester
Send message
Joined: 26 Dec 08
Posts: 132
ID: 33421
Credit: 12,510,712
RAC: 0
                
|
|
Is there any way to run fpsieve with more than 32 threads? Currently i have to run two instances :/
____________
 | |
|
HAmstyVolunteer tester
Send message
Joined: 26 Dec 08
Posts: 132
ID: 33421
Credit: 12,510,712
RAC: 0
|
|
for an amd 6276 is the performance drop severe.
2x -t16 2x83 Kp/s
1x -t32 1x200 Kp/s
that way i got for 2x -t32 only 2x 166 Kp/s, thats only 83% of the actual possible performance
____________
| |
|
Message boards :
Sieving :
Factorial Sieving discussion |
