General slow down in V25
 

I have several machines from PIII to CORE2 and they are all showing about 30% increase in interation times on the new version for LL. Sounds like some tuning might be in order.

[QUOTE=smoffat;150083]I have several machines from PIII to CORE2 and they are all showing about 30% increase in interation times on the new version for LL. Sounds like some tuning might be in order.[/QUOTE]What about the size of the numbers that you are dealing with? If you went to the next high FFT, you will se longer times, no matter what processor.

[QUOTE=Oleg V.Cat;147589]Strange, but it's real for my cheap desktop E2160*1.8Ghz. In one worker with one CPU thread I have iteration time approx ~0.090, with two - ~0.055 when CPU is free and ~0.090 in heavy loaded. In two workers I have approx ~0.140 for each worker thread...[/QUOTE]
Verify that Prime95 properly detects the processor capabilities. I had to do some manual worker affinity assignments on my E2180s; they were detected as single core hyperthreaded processors. Oddly, the E2200s work fine.

[QUOTE=Uncwilly;150093]What about the size of the numbers that you are dealing with? If you went to the next high FFT, you will se longer times, no matter what processor.[/QUOTE]

All the machines resumed work in progress with the new version of the software and the interation times all went up. I also noticed the software is using significantly more memory on all machines.

On the Core2 were you running one or 2 copies of Prime95? If you were running only 1 ov v24.x and now v25.x is running 2 tests, then what you are seeing is normal. If you have 2 worker threads (instead of the 1 for v24) then your through-put has gone up, but each test takes slightly long.

It might be too much memory.
 

I have a semaler problem keeping the CPU at a 100% so I turn down the memory use from 921 MBs to 850 MBs and the CPU runs great.
Note I have 1 Gig.

Hi,

[quote=smoffat;150117]All the machines resumed work in progress with the new version of the software and the interation times all went up. I also noticed the software is using significantly more memory on all machines.[/quote]
Did you set the affinity for each thread in the worker windows dialog?

And for me it's a little bit unclear what kind of configuration you had before and after switching to V25 on the DualCores.
V24 1 or 2 instances?
V25 1 or 2 instances? If 1, propper affinity settings?

[quote=smoffat;150117]
I have several machines from PIII to CORE2 and they are all showing about 30% increase in interation times on the new version for LL. Sounds like some tuning might be in order.
[/quote]
I'd just checked it with v25 and v24.14, both itteration times ar 47ms. (tested with one core only)

It's the last of the AMD 32bit one core I'm using.
 

I'm using soom of my scrap Computers to do the work and some are only single core and I set the memory low so it run at top speed.