rebirtherVolunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 2 Jan 13
Posts: 6956
Credit: 41,325,462
RAC: 58,268
by the new prst app. There are some issues with similar CPUs and Riesel Bases (-1), Sierpinski Base is fine:
GenuineIntel Intel(R) Celeron(R) J4125 CPU @ 2.00GHz [Family 6 Model 122 Stepping 8]
GenuineIntel Intel(R) Celeron(R) CPU N3150 @ 1.60GHz [Family 6 Model 76 Stepping 3]
GenuineIntel Intel(R) Pentium(R) Silver J5005 CPU @ 1.50GHz [Family 6 Model 122 Stepping 1]
The gwnum must have a bug, already reported to the dev.
If you have such errors with these CPUs, disable work for -1 bases until this is fixed.
/ 
Reply
Quote
I am glad it is just GWNUM (or the Intel CPU design itself?) that is at error here, it happened just when I added some extra libraries for multiple and arbitrary precision on my J5005....
BTW: The new Sierpinski Base app runs like a dream on one of my Ryzen 7 5700G machines, kudos.
P.S.: Get well soon!
/ Reply
Quote
rebirtherVolunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 2 Jan 13
Posts: 6956
Credit: 41,325,462
RAC: 58,268
I am glad it is just GWNUM (or the Intel CPU design itself?) that is at error here, it happened just when I added some extra libraries for multiple and arbitrary precision on my J5005....
BTW: The new Sierpinski Base app runs like a dream on one of my Ryzen 7 5700G machines, kudos.
P.S.: Get well soon!
must be something in gwnum lib, +1 bases are fine, something with FFT length I think.
/ Reply
Quote
by the new prst app. There are some issues with similar CPUs and Riesel Bases (-1), Sierpinski Base is fine:
GenuineIntel Intel(R) Celeron(R) J4125 CPU @ 2.00GHz [Family 6 Model 122 Stepping 8]
GenuineIntel Intel(R) Celeron(R) CPU N3150 @ 1.60GHz [Family 6 Model 76 Stepping 3]
GenuineIntel Intel(R) Pentium(R) Silver J5005 CPU @ 1.50GHz [Family 6 Model 122 Stepping 1]
1 (0x1) Unknown error number
The gwnum must have a bug, already reported to the dev.
If you have such errors with these CPUs, disable work for -1 bases until this is fixed.
Just a hunch: You do know that these CPU's have a thing in common? No L3 cache. If your code for -1 bases explicitely makes use of L3 cache than they will fail.
/ Reply
Quote
rebirtherVolunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 2 Jan 13
Posts: 6956
Credit: 41,325,462
RAC: 58,268
by the new prst app. There are some issues with similar CPUs and Riesel Bases (-1), Sierpinski Base is fine:
GenuineIntel Intel(R) Celeron(R) J4125 CPU @ 2.00GHz [Family 6 Model 122 Stepping 8]
GenuineIntel Intel(R) Celeron(R) CPU N3150 @ 1.60GHz [Family 6 Model 76 Stepping 3]
GenuineIntel Intel(R) Pentium(R) Silver J5005 CPU @ 1.50GHz [Family 6 Model 122 Stepping 1]
1 (0x1) Unknown error number
The gwnum must have a bug, already reported to the dev.
If you have such errors with these CPUs, disable work for -1 bases until this is fixed.
Just a hunch: You do know that these CPU's have a thing in common? No L3 cache. If your code for -1 bases explicitely makes use of L3 cache than they will fail.
Should work, +1 did, if Iam getting a chance to compile the latest prst source with gwnum 30.18 to test then I will know more, no feedback from the dev yet.
/ Reply
Quote
rebirtherVolunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 2 Jan 13
Posts: 6956
Credit: 41,325,462
RAC: 58,268
https://srbase.my-firewall.org/sr5/download/PRST-11-test.zip
This is a recompiled unchanged srbase version with latest code changes and gwnum30.19 for testing for the problem CPUs listed in this thread.
Run locally outside BOINC with
/ Reply
Quote