So who is going to be a guinea-pig this time??

Message boards : Number crunching : So who is going to be a guinea-pig this time??
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · Next

AuthorMessage
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1882924 - Posted: 10 Aug 2017, 4:24:49 UTC

I was wondering who ... if anyone .... is going to be a guinea pig for Threadripper on release tomorrow? The 1920X looks appealing and the mid-range motherboards look like they occupy a sweet spot. I especially like the fact that all the TR4 motherboards seem to support quad SLI or Crossfire-X with sensibly spaced X16 slots. Would have no issue putting 3 or 4 graphics cards in them to build a stupidly serious cruncher.

I have already started a Wish List on Newegg. The plan would be to replace at least one of my Win 7 crunchers.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1882924 · Report as offensive
Profile Bill G Special Project $75 donor
Avatar

Send message
Joined: 1 Jun 01
Posts: 1282
Credit: 187,688,550
RAC: 182
United States
Message 1882994 - Posted: 10 Aug 2017, 16:40:52 UTC - in response to Message 1882924.  

I expect to upgrade to Threadripper........but I have not figured out if the XATX board will fit in a ATX case and since there is no water cooling yet being made specifically for this chip, I am going to wait.
What I will be doing is to take down my farm and relying on one computer.
Amazon had it on preorder but now that it is shipping they are out of stock of both the chip and the MB.
There are so many innovations built into the MB that I am anxious to try them.
Release of necessary parts and finances will dictate when I do the upgrade.

SETI@home classic workunits 4,019
SETI@home classic CPU time 34,348 hours
ID: 1882994 · Report as offensive
Profile Darrell
Volunteer tester
Avatar

Send message
Joined: 14 Mar 03
Posts: 267
Credit: 1,418,681
RAC: 0
United States
Message 1883004 - Posted: 10 Aug 2017, 17:31:20 UTC - in response to Message 1882994.  

I expect to upgrade to Threadripper........but I have not figured out if the XATX board will fit in a ATX case and since there is no water cooling yet being made specifically for this chip, I am going to wait.
What I will be doing is to take down my farm and relying on one computer.
Amazon had it on preorder but now that it is shipping they are out of stock of both the chip and the MB.
There are so many innovations built into the MB that I am anxious to try them.
Release of necessary parts and finances will dictate when I do the upgrade.


List of compatible coolers from Guru3d review: http://www.guru3d.com/articles-pages/amd-ryzen-threadripper-1950x-review,5.html

And if you want to build your own custom cooling loop, I remember seeing a Thermaltake add pop up on my FB with blocks for the Threadripper.
... and still I fear, and still I dare not laugh at the Mad Man!

Queen - The Prophet's Song
ID: 1883004 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1883016 - Posted: 10 Aug 2017, 18:32:16 UTC - in response to Message 1882994.  

As long as your mid-size or full-size case doesn't have a step in the motherboard tray before the pass-through grommets, the EATX sized motherboards will fit if the case can handle ATX size.

The Threadripper SKUs all ship with an Asetek AIO bracket so if you have any of those coolers from Corsair etc. etc it will fit the chip. There are YT video showing how little of the active cores are actually covered by the round Asetek pump cold plate. Whether that is sufficient to cool the chip is the question. With the embargo lifted today, my task it to read through the reviews and see how that solution fairs.

There is going to be a full coverage air cooling solution from Noctua and I did in fact discover one product already announced that is a full coverage AIO solution from Enermax.
LIQTECH TR4 360

So there is at least one proper AIO cooling solution available now. Also there are announced full coverage water blocks from custom cooling companies like AlphaCool and their competitors.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1883016 · Report as offensive
Profile Bill G Special Project $75 donor
Avatar

Send message
Joined: 1 Jun 01
Posts: 1282
Credit: 187,688,550
RAC: 182
United States
Message 1883023 - Posted: 10 Aug 2017, 19:31:39 UTC - in response to Message 1883016.  

I knew there was something out there, but what I was seeing was that they were adapting the round cooler which did not cover the CPU. Thanks for the additional info.
I may have to bit the bullet sooner than later as I have one system that keeps rebooting one me. I think the first thing I will try next week is a new PS as all temps looks good.

SETI@home classic workunits 4,019
SETI@home classic CPU time 34,348 hours
ID: 1883023 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1883031 - Posted: 10 Aug 2017, 20:17:14 UTC

Just read through the Guru3D link posted in the thread. Apparently the partial coverage of the heat spreader with the Asetek round cold plate is a non-issue. And the fact that Threadripper is based on Ryzen core architecture means that all the legwork and development for the Zen processors with regard to BIOS, AGESA platform and memory compatibility has already been ironed out. I see that you are going to find easy running with Samsung B-die, single rank memory at the highest clocks and lowest latency. You just need four sticks instead of two for the quad channel configuration.

I can see that Threadripper is the new value entry for enthusiast, HEDT and content creation users.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1883031 · Report as offensive
Profile Shaggie76
Avatar

Send message
Joined: 9 Oct 09
Posts: 282
Credit: 271,858,118
RAC: 196
Canada
Message 1883059 - Posted: 10 Aug 2017, 23:00:55 UTC

I have a Haswell-E 8/16 5960X for my daily duties and the only dev benchmarks I've seen so far (Anandtech Chrome builds) were barely faster on ThreadRipper despite 2x the threads. Our builds regularly saturate all 16 threads and I so I assume Chrome does too so I'm suddenly much less interested in being an early adopter.
ID: 1883059 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1883075 - Posted: 11 Aug 2017, 0:02:41 UTC

A lot of test benchmarks I have read today show that there is still a lot of multi-thread benchmarks that are not using all available threads and the performance gap to Intel SKUs is minimal. The benchmarks that DO use all available threads are showing the expected performance gap between the 10core/20 thread parts and the TR 16core/32 thread 1950X. And the 1920X 12 core part is beating out the 1950X on single thread performance because of its higher base clock.

I am not sure if I will try out the architecture for crunching performance reasons and would only build it just for the curiosity of the new sub-systems. I do like the motherboards and could justify the mid-level products cost because of their connectivity options and the elimination of the gaming 'bling' that infests the high-end releases that I have no interest in. The Asus Prime X399-A and the Asrock X399 Taichi interest me. Those boards would allow fitting 3 GPUs without cooling compromises easily.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1883075 · Report as offensive
Al Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Avatar

Send message
Joined: 3 Apr 99
Posts: 1682
Credit: 477,343,364
RAC: 482
United States
Message 1883083 - Posted: 11 Aug 2017, 0:47:42 UTC - in response to Message 1883075.  

Did I hear that some of them are going to be releasing an E-ATX version of the board to support more GPUS with proper cooling clearances? Something about that sounds vaguely familiar.

ID: 1883083 · Report as offensive
Profile Wiggo
Avatar

Send message
Joined: 24 Jan 00
Posts: 34744
Credit: 261,360,520
RAC: 489
Australia
Message 1883091 - Posted: 11 Aug 2017, 1:44:29 UTC - in response to Message 1883083.  

Did I hear that some of them are going to be releasing an E-ATX version of the board to support more GPUS with proper cooling clearances? Something about that sounds vaguely familiar.

That would be the XL-ATX format for GPU clearances Al, but I'm waiting for the day when they combine both E-ATX and XL-ATX formats into the OBESE-ATX format. ;-)

Cheers.
ID: 1883091 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1883096 - Posted: 11 Aug 2017, 2:08:29 UTC - in response to Message 1883083.  

Did I hear that some of them are going to be releasing an E-ATX version of the board to support more GPUS with proper cooling clearances? Something about that sounds vaguely familiar.

All the ASUS sTR4 SKUs are E-ATX form factor.
amd-x399-motherboard-ryzen-threadripper-roundup-msi-asus-asrock-aorus

The Asrock, Gigabyte and MSI boards are ATX form factor and still offer 4 GPU installations with one slot clearance for ventilation input to the GPUs. But some of them put the M.2 NVMe ports under a hot GPU exhaust though. Take a look a the pictures and chart in the link in this post.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1883096 · Report as offensive
Ghia
Avatar

Send message
Joined: 7 Feb 17
Posts: 238
Credit: 28,911,438
RAC: 50
Norway
Message 1883136 - Posted: 11 Aug 2017, 7:39:35 UTC - in response to Message 1882994.  
Last modified: 11 Aug 2017, 7:40:18 UTC

I expect to upgrade to Threadripper........but I have not figured out if the XATX board will fit in a ATX case and since there is no water cooling yet being made specifically for this chip, I am going to wait.

EK has announced their new dedicated Threadripper water blocks.

[url]https://www.ekwb.com/news/ek-announcing-dedicated-amd-threadripper-supremacy-evo-water-blocks//url]
Humans may rule the world...but bacteria run it...
ID: 1883136 · Report as offensive
Al Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Avatar

Send message
Joined: 3 Apr 99
Posts: 1682
Credit: 477,343,364
RAC: 482
United States
Message 1883372 - Posted: 12 Aug 2017, 14:06:27 UTC

Ok, couple more questions, my X99 Asus was their rev. 1 board, and they came out with a 2nd one to update mine. I had tried using the m.2 slot when I bought it well over a year ago, and for the life of me couldn't get it to work. After doing a bit of research, and of course talking to Asus, I ended up returning it and going with SSD drives instead. I heard and read that the x99 wasn't really fully baked and not ready for prime time at the time of release, kinda like the X299 is today.

This being so, and the fact that there doesn't appear to be a compelling reason to move to the latest and greatest, my thoughts turned to, seeing they have now had a couple years to work things out, possibly getting a real solid X99 board instead. I spent a few late night hours searching for as current as possible comparative (Toms Hardware, etc) reviews of the most recently released X99 boards, and then started hitting the sites like Newegg and Amazon, to see how they fared in the real world.

I have to say, even knowing the fact that people who are having good experiences are much less likely to report such to those kind of sites, I was somewhat shocked at the things I was reading about a number of the boards, and pretty much from all the mfgs. I didn't see one that stood out head and shoulders above the rest in terms of features and especially reliability. I suppose I should take things with a grain of salt, but the best I can tell, even after a couple years and a couple revs, these still as a group don't seem to be rock solid (at least in the enthusiast class), and apparently the situation is, you pays your money, you takes your chances.

Bottom line takeaway from what I read is, if you are going to build one of these, do Not put it on the shelf and 'get around to it' later, because if there are issues with your specific unit, it's best to be well within the sellers 30 day return window, because otherwise you might end up with another persons (refurbished) problems in exchange for your full priced money spent, when you have to RMA it back to the Mfg. No thanks.

Anyone have thoughts/experiences with these latest X99 boards that they would be willing to share? It's pretty disheartening reading, spending between $3-500 on an enthusiast motherboard should get you something that is rock solid, and it appears that these are still not there. And I doubt there will be any more new ones released, as the X299 is now out, so what we have available now is as good as it's going to get.

ID: 1883372 · Report as offensive
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1883374 - Posted: 12 Aug 2017, 14:18:16 UTC - in response to Message 1883372.  

Apologies to the OP for the hijacking of this thread...

It's always been a buy beware sort of thing.

EVGA tends to be pretty good with RMA items so there is that. But they don't have a true 4 x16 pci board. But they seems to be the best (at least in my opinion)

That being said, I run ASUS boards (what??!!) Yes, you read that right. I do this because they have the X99e WS boards. The closest we have ever gotten to a true 4 X16 PCIe board. They have a new board out X99e 10ws which replaces their X99e 6ws that replaced the x99e ws usb3.1 that replaced.... you get the picture....

The problem with ASUS is customer service. It's horrible... Of the 5 boards I have bought from ASUS, 1 was DOA out of the box, the second died within a few months. When they work, they work great. But they don't always work..Really unsettling when you look at the cost of these boards.

After EVGA, ASUS, ASRock would be my 3rd choice. Gigabyte boards are ok, but not in the quality (in my opinion as the first 3)
ID: 1883374 · Report as offensive
Al Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Avatar

Send message
Joined: 3 Apr 99
Posts: 1682
Credit: 477,343,364
RAC: 482
United States
Message 1883381 - Posted: 12 Aug 2017, 15:46:50 UTC

Thanks for the thoughts, Zalster. I will check out EVGA again, and yes, amazingly, that has been the theme I have been reading about ASUS over the last number of years. Their support is Horrible, and my experience early last year validates that assessment. Good boards when they work though, apparently. Roll them dice...

And now back to our regularly scheduled Threadripper programming/discussion already in progress... ;-)

ID: 1883381 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1883392 - Posted: 12 Aug 2017, 17:02:54 UTC
Last modified: 12 Aug 2017, 17:40:37 UTC

I have only had to RMA two products so far and it was a 50-50 tossup. I thought I had a problem with my best GTX 970 and returned it to EVGA. I promptly got its replacement back in the solid black box and it seemed to not work in the same spot as the original. Further troubleshooting (which I should have done in the first place) proved the video card wasn't the problem. It was the ASUS motherboard. One of the PCIe slots had failed. Then I RMA'd the Sabertooth FX990 R2.0 back to ASUS. I received my exact same board back 3 months later with the same identical problem. They didn't fix or replace anything. And now I have a replacement GTX 970 which is a lot lower in ASIC quality and won't clock anywhere near my original, serial numbered sequentially and matched GTX 970s.

I did though replace the R2.0 Sabertooth with another ASUS product, the CHVFZ which has proven solid still. Don't know if I will have the same lousy RMA service experience with ASUS if I have another failure. I don't believe EVGA has ever made an AMD board but I would like to try them out at sometime.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1883392 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1883424 - Posted: 12 Aug 2017, 21:29:23 UTC - in response to Message 1883374.  

Apologies to the OP for the hijacking of this thread...

It's always been a buy beware sort of thing.

EVGA tends to be pretty good with RMA items so there is that. But they don't have a true 4 x16 pci board. But they seems to be the best (at least in my opinion)

That being said, I run ASUS boards (what??!!) Yes, you read that right. I do this because they have the X99e WS boards. The closest we have ever gotten to a true 4 X16 PCIe board. They have a new board out X99e 10ws which replaces their X99e 6ws that replaced the x99e ws usb3.1 that replaced.... you get the picture....

The problem with ASUS is customer service. It's horrible... Of the 5 boards I have bought from ASUS, 1 was DOA out of the box, the second died within a few months. When they work, they work great. But they don't always work..Really unsettling when you look at the cost of these boards.

After EVGA, ASUS, ASRock would be my 3rd choice. Gigabyte boards are ok, but not in the quality (in my opinion as the first 3)


I'm a happy runner of ASUS X99-E-WS and a 6/12 core Intel chip. The CPU runs 6 MB tasks and the 6 additional HT cores serve the GPUs that do the main work.

p.s. To run seti you do not need to have a full blown XXX8-16 * 4 PCIe configuration. The PCIe bus usage is about 1% - 4%. The SoG and my 'special' do their best at hiding the memory transfer latencies, i.e. they do other work a*) when the GPU to RAM memory transfer is occuring. The b*) RAM to GPU mem transfer is a one time event. *a) Happens often. It is hidden by launching work to the GPU before the memory transfer and doing CPU work when waiting for the memory transfer to complete. *b) Is run one time only.

p.p.s. Save your money to best available GPUs and the cheapest 4 core procesor and a MoBo with four 2x wide slots. Add a decent cooling and make sure your cards run near the performance limits set by the GPU hw and drivers. On NVIDIA that means overclocking the idle state since the GPU calculation work load runs a couple of performance states lower.

p.p.p.s I have had my ASUS MoBo replaced twice. The earlier models did not have (or I did not connect) the auxiliary PCIe power connector and I had the burnt up pin 22/23 or something problem.

Petri
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1883424 · Report as offensive
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1883426 - Posted: 12 Aug 2017, 21:38:23 UTC - in response to Message 1883424.  

I would only add that Petri's advice is SETI@HOME specific.

For those that run other projects, there is a need for the fastest PCIe available. I for one don't just run Seti. I run Einstein (whose new OpenCl is dependent since they don't have people that can fine tune the apps {heck, they need someone who understands wisdom kernals since they can't figure out theirs currently} and GPUGrid)

These other projects do vary throughput based on PCIe speeds. So take all of this with a grain of salt. My 2 cents
ID: 1883426 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1883431 - Posted: 12 Aug 2017, 21:57:28 UTC - in response to Message 1883426.  
Last modified: 12 Aug 2017, 21:59:32 UTC

A Good point Zalster. Thanks.
EDIT: and I stated that in the first p.s.
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1883431 · Report as offensive
Al Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Avatar

Send message
Joined: 3 Apr 99
Posts: 1682
Credit: 477,343,364
RAC: 482
United States
Message 1883439 - Posted: 12 Aug 2017, 22:35:08 UTC

Ok, I have to ask one more question, bouncing it off you guys for your thoughts, and then I'll either let this line of discussion die, or start a new thread. I want my CAD machine to be "future proof". Usually, in my way of thinking over the years, in the situation we find ourselves in right now, it'd be a pretty obvious choice for me. X99 is 'dying', X299 is in it's ascension, no brainier, right? Well, maybe, maybe not. The thought of buying/upgrading a system with a socket format that has no where else to go, vs. one that is just beginning, and hopefully has at least a few iterations of CPU's for it before that too goes into the scrapheap of IT history, well, it goes against the my better judgement to go with the dead end one. In fact, if I did do that, it would be the first time in well over 25 years of building PC's that I have. Getting one that is beginning vs. one that is (upgradeably) obsolete just goes against everything I've done in the past. But, that being said..

I really don't know. If I want to get this upgraded sooner rather than later, (and then again not even taking into account what I have heard that the latest procs are supposedly only supported on Win10, not 7) I'd be going with a 1.0 version of X299, which is pretty much what I did with X99, and that one didn't work out so well for me. Of course, I could just wait a few years till the 1.5 or 2.0 versions of X299 come out, but then we'd most likely be in the same situation, this time with the X1099 or whatever just around the corner. Typical PC conundrum, just wait a _little_ longer, the next version (that is going to be _so much_ better, you'll see!...) will be out next week/month/year, if you can just wait it out. Forever. Rinse and repeat. lol

From what I have been gathering, going with a X299 and a mid to upper end proc, (in terms of raw performance for my uses), isn't going to gain me much near term in performance over a top end X99, but the thing that makes it tougher is the fact that they would cost about the same amount, both the proc and the MB whichever way I went. The OS support is an issue if there are no workarounds on the newest chips for using Win7, but I guess it is the 1.0 thing that has me the most concerned. Go with a known quantity, verses one that might have issues, but will probably be able to be upgraded for cheaper and longer down the road. I have to think that I would be happy with the performance of the current X99 upgrade using fast U.2 drives and a pile of memory, it should perform admirably for at least a couple 3 years, and maybe by then, whatever is out can be financially justified by the decreased run times of whatever I am doing, and will pay for itself.

So, sorry for the ramblings, if you guys could add a couple quick thoughts to how I see things, I'd appreciate it. I really don't want to continue to monopolize his thread, but it might be something tangentially related to upgrading for people, even if I am talking about the X299 instead of the X399, but I think this part of the discussion could apply to both.

ID: 1883439 · Report as offensive
1 · 2 · 3 · Next

Message boards : Number crunching : So who is going to be a guinea-pig this time??


 
©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.