That config is pretty good (CPU+motherboard), case and power supply too, rest it crap. However, price is a bit too steep. You can buy lot of 2 E5-2683-V3 for as little as 700$ - http://www.ebay.com/itm/Lot-of-2-Intel- ... SwZQRYYf6g
Mobo (new one) is like 400-500$, case and power supply at best 100$. So the given config (and I am not counting memory and HDDs which virtually have 0 value) could be assembled for 1200-1300$, meaning those guys are selling it for double price.
Ryzen 1700 ($330) has 12-13% lower NPS than Ryzen 1800X ($500) and 8% lower NPS than Ryzen 1700X ($400) for Stockfish and Komodo. It represents the best buy in terms of Chess engine performance. Amazingly, Ryzen 1700 gives only 5% lower NPS than Intel 8-core i7-6900K at $1010 with yet unoptimized compiles. It gives +50% over Intel similarly priced ($340) 4-core i7-7700K. And +80% over my 4-core i7-4790, which I bought only a year ago. With Ryzen, speeds of 15-18 MNPS for Stockfish and Komodo will become common.
I have been out for a while ... then just saw the new AMD Rizen report on Chessbase website ... but I see you guys are way ahead of the game
Looks interesting ... especially the wattage used compared to the Intel! Has any optimizations been done on either SF or Komodo to better utilize the Rizen processors? Usually that tends to give a nice boost.
cma6 wrote:I ask that because my Intel dual-Xeon-E5-2686-V3 (total of 36 actual cores) is slower than my Intel i7-4930k (6 cores) for infinite analysis.
cma6 wrote:I ask that because my Intel dual-Xeon-E5-2686-V3 (total of 36 actual cores) is slower than my Intel i7-4930k (6 cores) for infinite analysis.
Hi Glenn:
My reason for my post was to hopefully find an expert on this board to help me diagnose and fix the problem with my dual-Xeon system. I have no interest in proving the speed to anyone.
Fortunately, one expert from this board took up the challenge and is working with me to diagnose the issue and fix it. In the first two days of analysis, we have checked CPU-Z screens and run DOS bench tests and have further work to do.
Thank you for your interest in the issue. However, I feel that I am in good hands with my expert.
Unfortunately, those are the things I can't tell you.
Schedules and specs are crazy sensitive. Official AMD press releases say the next generation will be out in the
1st quarter of 2018 and my team and me are working hard to ensure that. There are a lot of exciting things
coming from AMD but I can't really discuss too many of them (sadly). It's an exciting time to work here as an
analog circuit designer.
And all shilling and selfish reasons aside, I'm glad Intel has some competition again. It's almost enough to
make me start chess programming again!
Unfortunately, those are the things I can't tell you. :(
Schedules and specs are crazy sensitive. Official AMD press releases say the next generation will be out in the 1st quarter of 2018 and my team and me are working hard to ensure that. There are a lot of exciting things coming from AMD but I can't really discuss too many of them (sadly). It's an exciting time to work here as an analog circuit designer.
And all shilling and selfish reasons aside, I'm glad Intel has some competition again. It's almost enough to make me start chess programming again! :D
From what I've read you must be working on something beyond Naples. That's a surprisingly early release date if it doesn't slip.
