Message boards :
Number crunching :
Ryzen and Threadripper
Message board moderation
Previous · 1 . . . 7 · 8 · 9 · 10 · 11 · 12 · 13 . . . 69 · Next
Author | Message |
---|---|
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
That's the one I am eyeing too. The 12 and 16 core Ryzen 3000 cpus will have to suffice this year as the news today is that the Threadripper Gen3 cpus have been removed from this year's AMD roadmap. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
The new AMD Ryzen Gen 3 look interesting enough that I plan to buy one for a system I plan on building later this year. I'll wait for some good motherboards with PCIe 4.0 and perhaps that AMD will be friendly enough to release a newer Radeon VII that can utilize that bus. The CPU I'm eyeing is the Ryzen 7 3700X. And if they could release a MB with 6+ slots I would be all set ;) A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
I believe if that is ever to happen, it could/would only be possible with the new forthcoming X570 chipset motherboards. Haven't seen any technical details yet on the features of the X570 chipset or its architecture other than it is going to support PCIe 4.0. The MB OEM's have been holding those cards close to the vest. Haven't seen any leaks even. Guess we will have to wait for announcements at Computex next month. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
An inadvertent pre-product release of a BIOSTAR X570 motherboard was published today at Anandtech. https://www.anandtech.com/show/14322/the-biostar-x570-racing-gt8-motherboard-pcie-gen-4-ddr44000-and-triple-m2 Nothing too startling with respect to specs which closely follow the X470 chipsets other than a mention of DDR4-4000 memory speeds. That is a significant jump in DDR4 memory speeds from X470. Still only seeing two PCIe slots get supported from the cpu with an additional X4 slot coming from the PCH. So this board does not look like the answer for a AMD multi-gpu platform. [Edit] I found a better image of the board at wccftech.comhttps://wccftech.com/biostar-x570-racing-gt8-motherboard-amd-ryzen-3000-cpu-leak/ that shows the slots and the spacing better. You could theoretically install 3 gpus into the X16 slots and still mount risers into two of the three X1 slots that would be exposed and not covered by a two slot wide gpu. That could mean you could possibly support 5 gpus on this board. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Jord Send message Joined: 9 Jun 99 Posts: 15184 Credit: 4,362,181 RAC: 3 |
https://wccftech.com/amd-12-core-zen-2-ryzen-cpu-engineering-sample-benchmark-leaked-higher-ipc-vs-threadripper-1920x/ AMD 12 Core Zen 2 Ryzen 3000 Series CPU ES Benchmark Leaked – Higher IPC vs Threadripper 1920X So let’s get into the nitty gritty details. It’s crucial to first point out that this is an engineering sample, and one that we have in fact seen several months back. Although this time it’s been paired with a new motherboard, code named AMD Qogir-MTS. It features relatively low clock speed, as expected for an engineering sample. Running at a 3.4 GHz base clock and a 3.7 GHz Turbo. The production units are expected to operate at significantly higher clock speeds than that. The CPU has been paired with very modest 1600 MHz DDR4 memory and a 500GB HDD. Although, that’s not really the interesting part. The more interesting bit is how this chip compares to AMD’s existing 12 core Ryzen Threadripper 1920X. And it’s actually slightly faster, especially at floating point operations, despite having a slightly lower clock speed and significantly lower memory bandwidth. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
That's the one on my shortlist. I bet there will be little stock available and any appearing anywhere in the world will be instantly snatched up. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Interesting read over at Forbes speculating that Ryzen 3000 obviates need for Threadripper platform. TR may be killed off. https://www.forbes.com/sites/antonyleather/2019/05/15/amd-killing-off-threadripper-processors-suddenly-make-perfect-sense/#1ba9814556d1 Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
From the Tech Report's Ryzen 3000 rumours and leaks article, We expect AMD to announce the third-generation Ryzen CPUs at Computex starting on May 28, and hopefully products will actually launch by the time of E3 on June 11. Grant Darwin NT |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
All we can do at this point is hope that Dr. Su's keynote speech at Computex at the end of the month sheds more light on the product launches. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Interesting read over at Forbes speculating that Ryzen 3000 obviates need for Threadripper platform. TR may be killed off. I read something like that but I wonder. The "King of the Hill" for Workstations ideal it represents is significant. The EPYC has too low a slow cpu frequency to be a viable consumer replacement. If it is feasible to change the target socket to the EPYC socket so that the memory channel congestion issue goes away? Would that count has "backward" compatible? Nah.... but it would be a way to get past the memory channel congestion. I suppose "binning" a small percentage of 16 core and 32 core EPYC's for "very high" cpu frequencies and marketing them as Threadripper replacements would be another choice for the "King of the Hill" for Workstations ideal. Or simply keep selling the TR series and see what happens. The 2970wx/2990wx has some kind of niche that may not really go away. I think a Ryzen 3000 with 16c/32t would eat the bottom end of the Threadripper line just like the speculation thinks it would. Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
That's the development I expect. The 12 and 16 core TR's will disappear from the product stack leaving only the 24 and 32 core TR 3000's since the Ryzen 3700 and 3800 chips will suffice for the 12 and 16 core counts in the workstation niche. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
That's the development I expect. The 12 and 16 core TR's will disappear from the product stack leaving only the 24 and 32 core TR 3000's since the Ryzen 3700 and 3800 chips will suffice for the 12 and 16 core counts in the workstation niche. And cost less than the equivalent TR motherboards. So the other question is how hot will the 3700/3800 run? Will they also push you into the more expensive liquid cooling option or not? Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
The leaked specs on the 3700X is it still has the same 105W TDP of the 2700X. So would expect the same thermals. The 3800X has a TDP of 125W and the 3850X has a TDP of 135W. So even that chip is under the 140W TDP of my Intel i7-6850K Broadwell chip which I manage to keep around 60° C. overclocked to 4.250Ghz with a 280mm AIO. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Just stumbled onto a great image of the new X570 CPU/PCH PCIe layout in a OCN rumor thread. The OCN thread is here. https://www.overclock.net/forum/379-rumors-unconfirmed-articles/1725134-g3d-rumor-amd-seeds-board-partners-ryzen-3000-samples-runs-4-5-ghz-show-15-extra-ipc-14.html#post27973054 Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Just stumbled onto a great image of the new X570 CPU/PCH PCIe layout in a OCN rumor thread. That really looks like you could run 6 gpus just off the pcie slots. If you use the M.1 to gpu adaptors maybe 8. And who knows if you could run a 1 to 4 expander card and make it up past 9. I can see clearly now..... the rain (errr...upgrade) is gone..... (humming off pitch) Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
But you are going to have find compatible M.2 to gpu adapters that understand PCIe 4.0 speeds and same goes for PCIe 4.0 slot expanders. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
But you are going to have find compatible M.2 to gpu adapters that understand PCIe 4.0 speeds and same goes for PCIe 4.0 slot expanders. The 1 to 4 expanders (I have) apparently don't do Gen 3 but the individual adapters do. So the question maybe subject to experimental verification at best. And who knows if the riser base cards will run Gen 4. :( I clearly need a larger experimentation budget :) Tom A proud member of the OFA (Old Farts Association). |
ML1 Send message Joined: 25 Nov 01 Posts: 20291 Credit: 7,508,002 RAC: 20 |
Is this a new Threadripper killer?! All nodes lead to Rome: Epyc leak spills deets on second-gen Zen 32-core AMD... ... Back in March, we saw stats detailing the performance of the chip designer's upcoming 64-core, 128-thread flagship second-gen Epyc part. The info was revealed to the world by an anonymous user of SiSoftware's Sandra – a popular UK-made PC analysis and diagnostics tool that's been going strong since 1997 and includes the option to share benchmark results publicly online. The entry was subsequently deleted – but not before everyone managed to take a screenshot. Now, Sandra has outed the chip's 32-core, 64-thread sibling... Happy cool fast crunchin'! Martin See new freedom: Mageia Linux Take a look for yourself: Linux Format The Future is what We all make IT (GPLv3) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Is this a new Threadripper killer?! I think it is still too slow. Threadrippers have several things going for them. Lots of cores, lots pcie lanes and speeds similar to the 2700x on the AM4 socket. When you offer the first two advantages but a cpu turbo that is still below 3Ghz it doesn't make sense that it would be a 2970x/2990x killer. Unless you are betting that high powered multiple gpu's would over ride the really low cpu speed(s). I wonder if the chip maker would be able to manufacture the "older" TR chips without limiting the 7nm ramp up. If yes, it would make perfect sense to continue to sell Threadrippers while you see if the latest EYPC version would start eating into the 2970x/2990x sales. Tom A proud member of the OFA (Old Farts Association). |
ML1 Send message Joined: 25 Nov 01 Posts: 20291 Credit: 7,508,002 RAC: 20 |
The suspicion is that those first numbers are for Engineering Samples... So who can guess what numbers for what comes out in production when finally announced/released? Good and interesting that development continues apace and looks to be giving Intel a very hard run... Especially so when you factor in the mitigations needed to run the Intel CPUs: Intel Loses 5X More Average Performance Than AMD From Mitigations: Report AMD Immune to Crippling MDS Vulnerabilities Interesting times... Happy fast crunchin', Martin See new freedom: Mageia Linux Take a look for yourself: Linux Format The Future is what We all make IT (GPLv3) |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.