mmt Vs. Ovyron (G4 D5 BG2)

[SheikhYerbouti]

Aha, so I missed Nc5. But if the reason was that black had some move that was killing 22.Bh3 that I couldn't see (it falls below Nc5), this is actually good. In that case I'd have played 22.Bh3 if I was white and lost faster. Let's see if I can reach the -1.90s now...

So what you're saying is [in Cathy Newman voice] SF never saw Nc5 as preferable to Bh3, or you didn't? How many multi PVs do we need. Whose nose is out of joint?

[jp]

1r1qr1k1/p3bpp1/2p1bnn1/2Np2Bp/3P3P/1P1QPP2/P1N3B1/R4K1R b - - 1 1

Komodo 13.2, depth 28: -1.11 (... Bc8).

At depth 33, Komodo's top three moves (which I won't name here) have evals -0.83, -0.72, -0.53.

[mmt]

Initial tests with the new drive show that it can sustain more than 3 times what my other drive can. Testing them is tricky. It's hard to find positions that will drive the TB requests to high numbers and keep them their long enough so that you can go to the right screens and record the data. Finding a position that can drive the request high enough for long enough to cause the drive to heat up to the point of throttling seems problematic. But this is a good thing I think. I haven't been able to test its peak yet but it sustains about 16k+ hard page faults a second and over 1.2GB of random reads per second with ease. My other drive could sustain 4k-5k and about 200Mb of random reads. This allow SF to maintain about 80-90 Mnps while heavy requests to 7-man TB files. I would say it's worth the price premium over a "standard" NVMe drive. One other advantage is it has is 8 times the write life. Not sure I need this since TB request are read only operations but I do have apps that beat the hell out of my hard drives with constant read/write cycles. This would be a good drive to run those apps. Now I just need 4 x 4TB drives like this one in a striped array and I'll be good to go!

The NVMe drives are great but they only go up to 2TB and I only have one spot for them on the motherboard. I'm keeping all the 7-piece TBs on regular SSDs (need the NVMe drive for other things) but I'm pretty much out of space on them. That's a good number of nodes/s for SF!

[jp]

Komodo 13.2, depth 35: -1.32.
It's eval got worse in the last two plies (see post above).

[Zenmastur]

Initial tests with the new drive show that it can sustain more than 3 times what my other drive can. Testing them is tricky. It's hard to find positions that will drive the TB requests to high numbers and keep them their long enough so that you can go to the right screens and record the data. Finding a position that can drive the request high enough for long enough to cause the drive to heat up to the point of throttling seems problematic. But this is a good thing I think. I haven't been able to test its peak yet but it sustains about 16k+ hard page faults a second and over 1.2GB of random reads per second with ease. My other drive could sustain 4k-5k and about 200Mb of random reads. This allow SF to maintain about 80-90 Mnps while heavy requests to 7-man TB files. I would say it's worth the price premium over a "standard" NVMe drive. One other advantage is it has is 8 times the write life. Not sure I need this since TB request are read only operations but I do have apps that beat the hell out of my hard drives with constant read/write cycles. This would be a good drive to run those apps. Now I just need 4 x 4TB drives like this one in a striped array and I'll be good to go!

The NVMe drives are great but they only go up to 2TB and I only have one spot for them on the motherboard. I'm keeping all the 7-piece TBs on regular SSDs (need the NVMe drive for other things) but I'm pretty much out of space on them. That's a good number of nodes/s for SF!

I've been testing TBs with different configurations of memory free ect. My preliminary conclusions is you shouldn't use 7-man TB for long (overnight) analysis unless you have enough memory to store all the TB files that will be needed in memory OR you have an exceedingly fast NVMe raid array. The reason is that once memory fills up you will be dependent on the SSD to swap files in and out of memory. This slows the node processing rate to 25% of its normal value on my machine even with a fast NVMe drive. I set two versions of SF running on the same position. One with 7-man with very little memory left over and only 6-threads it slowed to 2.7Mnps when the memory got full while the other version was running without table bases was maintaining 14.8Mnps also on 6 threads. I let them run at the same time. The one without TB got to depth 67 in less than 30 minutes. With 7-man TBs it took 40 minutes to get to depth 50. It might be better with a raid array. But those tests will have to wait for another time.

When you have enough memory this doesn't seem to slow the CPU down by much. I'm going to test that next. But I'm going to use 6-man TB because I know I can let it run for a fair amount of time without the fear of filling the memory. Plus JP and others, myself included, would like to know what effect 6-man have on long time control analysis.

Regards,

Zenmastur

[mmt]

Thanks for the info Zenmaster, interesting. I do have a lot of RAM, 512GB.

[Ovyron]

So what you're saying is [in Cathy Newman voice] SF never saw Nc5 as preferable to Bh3, or you didn't? How many multi PVs do we need. Whose nose is out of joint?

I did check Nc5 long time ago (back when the position first appeared in analysis), but discarded it as a much inferior choice than Bh3 (because I could never find the moves that beat Bh3). I'm confident I can drag the scores up to some -2.17, and prove Be6 was better, but after Zenmastur claimed he could drag the score to -3.00, that's not saying much.

1. g4 d5 2. Bg2 Bxg4 3. c4 c6 4. Qb3 e6 5. Qxb7 Nd7 6. Nc3 Ne7 7. cxd5 exd5 8. d4 Rb8 9. Qa6 Rb6 10. Qd3 Ng6 11. h3 Be6 12. Nf3 Bd6 13. h4 h5 14. b3 Nf6 15. Bg5 O-O 16. e3 Re8 17. Kf1 Bg4 18. Ne1 Bb4 19. Na4 Rb8 20. Nc2 Be7 21. f3 Be6 22. Nc5 Bc8

[d]1rbqr1k1/p3bpp1/2p2nn1/2Np2Bp/3P3P/1P1QPP2/P1N3B1/R4K1R w - -

[zullil]

1. g4 d5 2. Bg2 Bxg4 3. c4 c6 4. Qb3 e6 5. Qxb7 Nd7 6. Nc3 Ne7 7. cxd5 exd5 8. d4 Rb8 9. Qa6 Rb6 10. Qd3 Ng6 11. h3 Be6 12. Nf3 Bd6 13. h4 h5 14. b3 Nf6 15. Bg5 O-O 16. e3 Re8 17. Kf1 Bg4 18. Ne1 Bb4 19. Na4 Rb8 20. Nc2 Be7 21. f3 Be6 22. Nc5

[d]1r1qr1k1/p3bpp1/2p1bnn1/2Np2Bp/3P3P/1P1QPP2/P1N3B1/R4K1R b - - 1 1

Komodo at depth 54 had 22. Bh3 (-1.36).

[mmt]

1. g4 d5 2. Bg2 Bxg4 3. c4 c6 4. Qb3 e6 5. Qxb7 Nd7 6. Nc3 Ne7 7. cxd5 exd5 8. d4 Rb8 9. Qa6 Rb6 10. Qd3 Ng6 11. h3 Be6 12. Nf3 Bd6 13. h4 h5 14. b3 Nf6 15. Bg5 O-O 16. e3 Re8 17. Kf1 Bg4 18. Ne1 Bb4 19. Na4 Rb8 20. Nc2 Be7 21. f3 Be6 22. Nc5 Bc8 23. Kf2

[d]1rbqr1k1/p3bpp1/2p2nn1/2Np2Bp/3P3P/1P1QPP2/P1N2KB1/R6R b - - 1 1

[zullil]

1. g4 d5 2. Bg2 Bxg4 3. c4 c6 4. Qb3 e6 5. Qxb7 Nd7 6. Nc3 Ne7 7. cxd5 exd5 8. d4 Rb8 9. Qa6 Rb6 10. Qd3 Ng6 11. h3 Be6 12. Nf3 Bd6 13. h4 h5 14. b3 Nf6 15. Bg5 O-O 16. e3 Re8 17. Kf1 Bg4 18. Ne1 Bb4 19. Na4 Rb8 20. Nc2 Be7 21. f3 Be6 22. Nc5 Bc8 23. Kf2

[d]1rbqr1k1/p3bpp1/2p2nn1/2Np2Bp/3P3P/1P1QPP2/P1N2KB1/R6R b - - 1 1

So far, Komodo remain unimpressed by 22. Nc5. At depth 51 it had 22...Bc8 23. Kf2 (-1.45).