|
Message boards :
Sieving :
GPU slow, carry flag warning
| Author |
Message |
|---|
|
|
|
Hello everybody,
Is a Sapphire Radeon Pro Duo slower than an
INNO3D GeForce GTX 980 iCHILL HerculeZ X3 Air Boss ?
Sapphire Radeon Pro Duo ca.49.000/s (26,5P/day)
INNO3D GeForce GTX 980 iCHILL ca.133.500/s (69,5P/day)
start with
C:\GFN17\gfnsvocl_w64_2G 17 156570 156690
For the ATI GPUs I get this error message
C:\Users\MAX_29~1\AppData\Local\Temp\\OCL4016T3.cl:70:2:
warning: No built-in carry flag support known, using plain C code
#warning No built-in carry flag support known, using plain C code
^
1 warning generated.
There can be mine PCI Express 3.0 x16 nVIDIA GeForce GT 650M (Acer) ca. 25.450/s (13.5P/day) keep up.
Why is the Sapphire Radeon Pro Duo so slow ?
Greetings from Germany
Thorsten | |
|
 RafaelVolunteer tester
Send message
Joined: 22 Oct 14
Posts: 912
ID: 370496
Credit: 551,870,004
RAC: 455,383
                         
|
Why is the Sapphire Radeon Pro Duo so slow ?
Greetings from Germany
Thorsten
It's partially because you're not using the B(1-13) argument. Lower B have lower performance, whereas higher B offer A LOT more performance (up to a point, depending on the card) at the cost of screen lag. The app uses B7 as the default, but you can play with it to squeeze more out of your card. Try B9 or B10 for both cards, see how that goes. | |
|
|
|
|
Hee thank you for the quick response
C:\GFN17\gfnsvocl_w64_2G 17 156690 156691 B10
Result: ca. 149.500 (78,5P/day)
Thank you very much | |
|
|
|
Hee thank you for the quick response
C:\GFN17\gfnsvocl_w64_2G 17 156690 156691 B10
Result: ca. 149.500 (78,5P/day)
Thank you very much
Don't forget to use the W1 parameter also which will reduce CPU usage. Works on each GPU I've tried it on where it could.
| |
|
RafaelVolunteer tester
Send message
Joined: 22 Oct 14
Posts: 912
ID: 370496
Credit: 551,870,004
RAC: 455,383
|
Hee thank you for the quick response
C:\GFN17\gfnsvocl_w64_2G 17 156690 156691 B10
Result: ca. 149.500 (78,5P/day)
Thank you very much
Don't forget to use the W1 parameter also which will reduce CPU usage. Works on each GPU I've tried it on where it could.
*For Nvidia cards only.
**For "high" N values, it won't do anything for 15~17 (at the very least). | |
|
|
|
Hee thank you for the quick response
C:\GFN17\gfnsvocl_w64_2G 17 156690 156691 B10
Result: ca. 149.500 (78,5P/day)
Thank you very much
Don't forget to use the W1 parameter also which will reduce CPU usage. Works on each GPU I've tried it on where it could.
*For Nvidia cards only.
**For "high" N values, it won't do anything for 15~17 (at the very least).
I know it says Nvidia only but it also works for my Intel. And yes about the N values. Those use more CPU regardless of the w1 parameter or not. I did see a slight improvement for GFN18, those it was minor compared to GFN20-GFN22. The faster your CPU the less CPU time it uses from what I have seen. My Intel cores use less % of CPU then my AMD cores across the board whether its with the w1 parameter or not. | |
|
Crun-chi Volunteer tester
Send message
Joined: 25 Nov 09
Posts: 3233
ID: 50683
Credit: 151,443,349
RAC: 54,956
                      
|
Build Log (211 bytes):
C:\Users\Omega\AppData\Local\Temp\\OCL2460T3.cl:70:2: warning: No built-in carry flag support known, using plain C code
#warning No built-in carry flag support known, using plain C code
^
1 warning generated.
I got same error but I am using B value.
Also I got same error on Windows 7 or 10, same error using Radeon 290X or Radeon 580X
And last , maybe important : that error doesnot show until you reach around 8000P as start point for sieving, anything above throw error: Everything below works perfect.
____________
92*10^1585996-1 NEAR-REPDIGIT PRIME :) :) :)
4 * 650^498101-1 CRUS PRIME
2022202116^131072+1 GENERALIZED FERMAT
Proud member of team Aggie The Pew. Go Aggie! | |
|
Message boards :
Sieving :
GPU slow, carry flag warning |
