Search found 189 matches

by dragontamer5788
Fri Dec 20, 2019 4:47 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: I bought a stop-gap machine
Replies: 34
Views: 4072

Re: I bought a stop-gap machine

Good point , the 3950x can be found for $900 at retail and the 10980Xe is currently retailing at $1100. They are pretty close in value. If you have custom built system by Puegot, the 3950x system will be $500 less than 10980xe, currently $2600 with Gt1030 and 16Gb ram, 1TB 7200 SATA - but most folk...
by dragontamer5788
Fri Dec 20, 2019 3:48 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: I bought a stop-gap machine
Replies: 34
Views: 4072

Re: I bought a stop-gap machine

Projected shipping date is 12/12/2019. I believe that the projected delivery date was the 20th. wow - I just got my delivery date from Velocity - 12/23. Earlier, that was my "to be shipped date". A modernized Baby Blue is coming home for Christmas - less than 96 hours. https://www.dropbox.com/s/pi6...
by dragontamer5788
Tue Dec 17, 2019 5:34 pm
Forum: Computer Chess Club: General Topics
Topic: AMD 5700 XT and 5700 giving Flaky results to SETI@Home
Replies: 3
Views: 566

Re: AMD 5700 XT and 5700 giving Flaky results to SETI@Home

Glad I bought Nvidia Well, at your price range, you probably would have been getting Radeon VII instead. NAVI as an ISA looks excellent overall, but without ROCm support, and with this new OpenCL issue coming up, I have to say that any compute-oriented person probably should stay away and stick to ...
by dragontamer5788
Tue Dec 17, 2019 4:20 am
Forum: Computer Chess Club: General Topics
Topic: AMD 5700 XT and 5700 giving Flaky results to SETI@Home
Replies: 3
Views: 566

AMD 5700 XT and 5700 giving Flaky results to SETI@Home

https://setiathome.berkeley.edu/forum_thread.php?id=84508&sort_style=6&start=0 Seems like the AMD Navi GPUs aren't yet solid enough for SETI@Home. AMD Fans probably should stick with Vega a bit longer if you're doing GPU compute. Reddit Thread: https://www.reddit.com/r/BOINC/comments/ebiz18/psa_plea...
by dragontamer5788
Mon Dec 16, 2019 4:26 am
Forum: Computer Chess Club: General Topics
Topic: The Mac Pro is back - finally
Replies: 19
Views: 3895

Re: The Mac Pro is back - finally

...obviously the new Mac Pro does not play in the same NPS/Dollar league like systems based on recent AMD Zen CPUs for Stockfish, or Nvidia RTX series for LC0, but for Video/CGI the quad-GPU deal seems a good one, you get 4 AMD Vega II GPUs (with each 32 GB HBM) for ~10K $. -- Srdja The main issue ...
by dragontamer5788
Wed Dec 11, 2019 5:53 pm
Forum: Computer Chess Club: General Topics
Topic: The Mac Pro is back - finally
Replies: 19
Views: 3895

Re: The Mac Pro is back - finally

A major advantage to the Mac Pro is that its consistently a Skylake-X CPU + AMD Vega GPU. Anyone writing programs for a Mac Pro can very specifically tailor their application for that combination. In particular, we know that SkylakeX has PDEP, PEXT, and AVX512 implemented. AMD Vega GPUs have differe...
by dragontamer5788
Wed Dec 11, 2019 4:23 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: PEXT/PDEP are even slower than you think on Zen
Replies: 12
Views: 2197

Re: PEXT/PDEP are even slower than you think on Zen

At the moment I do a second calculation to get passed-pawn locations that I need to calculate their interactions with pieces Passed Pawn actually seems like an ideal candidate for pdep. Pull all the bits where pawns could block. Ex: White has pawn on E4. Check for black pawns on D5, E5, F5, D6, E6 ...
by dragontamer5788
Wed Dec 11, 2019 4:23 am
Forum: Computer Chess Club: General Topics
Topic: BootChess (minimal chess engine)
Replies: 38
Views: 8358

Re: BootChess (minimal chess engine)

mar wrote:
Fri Jan 30, 2015 12:29 am
Actually the program could always resign, making it even smaller (17 bytes):

Code: Select all

org 100h
mov ah, 9
mov dx, msg
int 21h
retn

msg db 'I resign$'
But yours is clearly stronger.
Mine is even shorter. Mine will wait out the clock, losing by time-out. The code is as follows:

Code: Select all

jmp $
by dragontamer5788
Tue Dec 10, 2019 9:09 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: PEXT/PDEP are even slower than you think on Zen
Replies: 12
Views: 2197

Re: PEXT/PDEP are even slower than you think on Zen

At the moment I do a second calculation to get passed-pawn locations that I need to calculate their interactions with pieces Passed Pawn actually seems like an ideal candidate for pdep. Pull all the bits where pawns could block. Ex: White has pawn on E4. Check for black pawns on D5, E5, F5, D6, E6 ...
by dragontamer5788
Tue Dec 10, 2019 7:21 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: PEXT/PDEP are even slower than you think on Zen
Replies: 12
Views: 2197

Re: PEXT/PDEP are even slower than you think on Zen

Another consideration is noise, the 180W thermals from the 10980XE can be cooled down very quietly with a Noctua NH-D15, I have two systems with these running and I can hardly hear that they are turned on. For a TR3-3970X you will start at 280W (probably even more), a 64 core TR3 will be like 500W,...