Laskos wrote:
I re-booted, all was clean, Win 7 installation is new, I don't know what might went wrong.
LOL.
I didn't imply something got "wrong". Actually you should be happy about those results.
Probably the 2600 is (much) better for HT than my humble (as of now) 9200 (that is beginning to get old).
I wonder at this point what something like a 3930k (or similar) would be capable of, and what the difference between HT and not is with 6 threads instead than 4.
I myself was surprised. My percentages are (nps8/nps4) -1, I guess you mean the same. Anyway, very high values, curious about those 3930k and 3960X beasts.
I am more positive now that HT is beneficial on my comp. I have a collection of 100 hard positions, and even time to solutions was lower now with 8 threads. Averaged results in 3 runs:
Tactical Mode, 20 seconds/position, 1GB Hash
8 threads: 65/100, 4.8s to solution
4 threads: 58/100, 5.2s to solution
I guess a healthy 10-20 Elo points from HT.
And the last test, pretty convincing from my point of view. I took "Hard test set talkchesscom-2012" by Vincent Lejeune http://talkchess.com/forum/viewtopic.php?t=44531 consisting of 218 hard positions.
Tactical Mode, Hash 2GB, time: 60s/position
4 threads: 131/218 Average time: 11.04s
8 threads: 142/218 Average time: 10.05s
Pretty conclusive, not only significantly more solutions with 8 threads, but 10% faster to solutions. HT with its 27-36% speed-up on my comp seems clearly beneficial.
Laskos wrote:Pretty conclusive, not only significantly more solutions with 8 threads, but 10% faster to solutions. HT with its 27-36% speed-up on my comp seems clearly beneficial.
Kai
Thank you for the results. Your data are very much in line with the expectation that any HT speed gain above 20% provides a real strength improvement.
It is quite possible that this only occurs with the more recent Intel CPU architectures: Sandy Bridge (i7-2600/2700/3930/3960) and Ivy Bridge (i7-3770). Possibly the larger speed gains only occur with the quad CPUs, hard to tell without further data.
Maybe someone can make the test with Houdini 3 on a i7-3770K.
Laskos wrote:And the last test, pretty convincing from my point of view. I took "Hard test set talkchesscom-2012" by Vincent Lejeune http://talkchess.com/forum/viewtopic.php?t=44531 consisting of 218 hard positions.
Tactical Mode, Hash 2GB, time: 60s/position
4 threads: 131/218 Average time: 11.04s
8 threads: 142/218 Average time: 10.05s
Pretty conclusive, not only significantly more solutions with 8 threads, but 10% faster to solutions. HT with its 27-36% speed-up on my comp seems clearly beneficial.
Kai
Nice news, Houdini 3 again +20 elo stronger !
BTW : please, send me the log of the test suite (the 2 run, if possible) via private message or in a new thread here.
Vinvin wrote:Nice news, Houdini 3 again +20 elo stronger !
Oops, +10% means +7 Elo, not +20 ...
A bit more than 7 points, probably 10-20. It is a double effect: more solutions found and in shorter time.
Kai
One cannot reliably extrapolate tactical test suite results to Elo strength.
Adding more threads will enlarge the width of the search - great for tactical puzzles but not necessarily for real game play.
the significant improvement in hyperthreading has been known for some time...
Microsoft worked very closely w/Intel during the development of WIndows 7, and significantly beefed up support for HT.
a large proportion of the best engines on playchess are now running w/ hyperthreading enabled,
why?, the operators noticed a significant performance improvement
there's no Houdini 'magic' here...regardless of what the author may want you to believe...
all SMP engines should benefit equally (if run on a modern Intel architecture w/ Windows 7)
Houdini: 0 net ELO gain when tested against other SMP opponents on same architecture
kranium wrote:
Houdini: 0 net ELO gain when tested against other SMP opponents on same architecture
How many thousands games did you play to get an error bar < 1 ?
The fact that H2 apparently only gains 8% by HT already disproves his theory. Different engines benefit to different degrees from HT. Add to that the fact that different engines incur different amounts of search overhead when the number of threads doubles.
Laskos wrote:And the last test, pretty convincing from my point of view. I took "Hard test set talkchesscom-2012" by Vincent Lejeune http://talkchess.com/forum/viewtopic.php?t=44531 consisting of 218 hard positions.
Tactical Mode, Hash 2GB, time: 60s/position
4 threads: 131/218 Average time: 11.04s
8 threads: 142/218 Average time: 10.05s
Pretty conclusive, not only significantly more solutions with 8 threads, but 10% faster to solutions. HT with its 27-36% speed-up on my comp seems clearly beneficial.
Kai
As tactical suite is not very accurate to estimate the strength of an engine, I've another system to evaluate the speed : In the starting position, with 8 best lines, measure time to reach depth 22 (it's about 1 min on your PC), make this operation 10 times with and without hyper-threading, make the average for both and report the time here ...
