Superpi 32m (Spi32m) final loop duration

[Stelaras]

Here's a quick and easy question ...

 

What are the most importand factors that prescribe the last loop duration ??? (After the 24th loop) .

 

 

PS . I hope i used the right word .

Edited August 3, 2010 by Stelaras

[TaPaKaH]

properly made waza is very important for last loops (20-24), if you manage to get the available and sys.cache equal or very close (diff <100) with both being high, you can drop a good second or two there

[Stelaras]

properly made waza is very important for last loops (20-24), if you manage to get the available and sys.cache equal or very close (diff <100) with both being high, you can drop a good second or two there

 

 

OK . Got it .

 

You mean one second or two in the last loops (20-24) or only in the final loop ???

[Stelaras]

OK . I managed to gain around 2secs with copywaza . That's for the whole spi32m run . So i guess that's what you were talking about .

 

The final loop gained only something like 0.1-0.2secs .

 

Is there something else that can give a nice boost to the final loop ??

Maybe a faster harddrive , ram drive , SSD ... etc etc.

Edited August 5, 2010 by Stelaras

[BenchZowner]

Use FAT32 filesystem on both partitions ( OS/SuperPi & CopyWaza ).

Fixed-size pagefile on SSD/Acard should help ( not sure though ).

[Stelaras]

I always use FAT32 on both partitions/disks . Pagefile always locked at 512MB .

 

I tried ... HDD 7200RPM , HDD 10000RPM , SSD Intel Postville ... And Acard .

 

No differences at all . Less than 0.1sec for sure (Lucky runs) .

 

So ??? What else gives boost to the final loop ??

Is there anyone that can comment ???

Edited August 14, 2010 by Stelaras