Is this normal

wayne lowrance · Post by **wayne lowrance** » Thu Jul 10, 2008 4:21 pm

Howdy, Just purchased H12. I like it. Have aquestion.
I set it to infinate analysis over nite, when I went to bed the Np/s was several thousand when I got up it was under 100. My comp was working fine. All 8 cores 100 %. Clock speed normal, FSB normal, core temp normal for all 8 cores running 100%. When I entered the first move in final pv, lthe nodes went back up to severa thousand.
Normal ?

thank you
Wayne

Post by **Harvey Williamson** » Thu Jul 10, 2008 4:31 pm

wayne lowrance wrote:Howdy, Just purchased H12. I like it. Have aquestion.
I set it to infinate analysis over nite, when I went to bed the Np/s was several thousand when I got up it was under 100. My comp was working fine. All 8 cores 100 %. Clock speed normal, FSB normal, core temp normal for all 8 cores running 100%. When I entered the first move in final pv, lthe nodes went back up to severa thousand.
Normal ?

thank you
Wayne

Yup - everything is ok - most engines do this. As long as its running at 100% cpu all is well.

Vempele · Post by **Vempele** » Thu Jul 10, 2008 4:33 pm

I'm guessing the node count overflowed. It should take about 2 or 4 billion nodes.

All it takes to fix it is to change one line of code. So if that's the case:

Post by **Harvey Williamson** » Thu Jul 10, 2008 4:38 pm

Vempele wrote:I'm guessing the node count overflowed. It should take about 2 or 4 billion nodes.

All it takes to fix it is to change one line of code. So if that's the case:

At least hiarcs does not subtract 3 from its search depth and divide its nodes per second by about 10

Vempele · Post by **Vempele** » Thu Jul 10, 2008 4:41 pm

Harvey Williamson wrote:
Vempele wrote:I'm guessing the node count overflowed. It should take about 2 or 4 billion nodes.

All it takes to fix it is to change one line of code. So if that's the case:

At least hiarcs does not subtract 3 from its search depth and divide its nodes per second by about 10

Where's the STRELARCS source code to prove it?

Post by **Harvey Williamson** » Thu Jul 10, 2008 4:51 pm

Vempele wrote:
Harvey Williamson wrote:
Vempele wrote:I'm guessing the node count overflowed. It should take about 2 or 4 billion nodes.

All it takes to fix it is to change one line of code. So if that's the case:

At least hiarcs does not subtract 3 from its search depth and divide its nodes per second by about 10
Where's the STRELARCS source code to prove it?

That source code wont be allowed to be posted here

HIARCS Chess Forums

Is this normal

Is this normal

Re: Is this normal