Author Topic: Threadripper Builds  (Read 31110 times)

2019-02-26, 04:04:50
Reply #285

dfcorona

  • Active Users
  • **
  • Posts: 113
    • View Profile
What do you guys think of the MASTERLIQUID ML360 RGB TR4 EDITION for cooling?

2019-02-27, 07:39:57
Reply #286

Vuk

  • Active Users
  • **
  • Posts: 42
    • View Profile
I have read on some reviews that its poor build quality and made of plastic which is pron to brake after a while. Haven’t used it myself though...

2019-02-27, 08:13:45
Reply #287

dfcorona

  • Active Users
  • **
  • Posts: 113
    • View Profile
What do you guys think of the MASTERLIQUID ML360 RGB TR4 EDITION for cooling?
wow, yes you are right. Total junk, they use plastic for the mounting onto the CPU which seems to break for most. Damn there really isn't a good AIO out there for 2990wx.
« Last Edit: 2019-02-27, 09:01:49 by dfcorona »

2019-02-28, 08:01:25
Reply #288

Vuk

  • Active Users
  • **
  • Posts: 42
    • View Profile
Personally I would go for a Enermax liqtech edition 2 360 or 280. There is the whole story of edition 1 having leaking issues and seems that even some people have this problem with the revision 2. I do own an edition 1 280 for more then 14 months now on a 1950x and thus far no issues at all. So its a loterry... But seems that all problems start within the warranty period so I guess if you have any you could solve them with that. Thats the only viable budget option if you want “serious” overclocking. The other route is custom loop so prepare for 400-700 euro expenses down the line.

I decided to sit out the whole TR2 generation and see what june/july 2019 brings up. I guess new boards and coolers will be introduced then for Tr3/ryzen 3000 and there will be more new aio’s.

2019-03-13, 09:35:25
Reply #289

tallbox

  • Active Users
  • **
  • Posts: 96
  • George Nicola
    • View Profile
    • Architectural Visualisations TALLBOX
Hi guys,
I am looking to get a new single slave for my day to day renderings. Which one would work better with max-corona
I have a budget of 4ooo euros. I have 1080 already so I don't need to buy one.

PS: I have read all the previous pages, so the clear winner is 2990wx. Thanks for the great thread
« Last Edit: 2019-03-16, 13:32:18 by tallbox »
Architectural Story Telling / Deep work practitioner
https://www.tallbox.co.uk

2019-03-20, 20:24:22
Reply #290

DUMNY

  • Users
  • *
  • Posts: 3
    • View Profile
Hello, all corona freaks

For several weeks I have been working on a new machine that I have made myself. The rendering engines I mainly use are v-ray and of course corona. Vray benchmark definitely satisfies me, unfortunately the corona benchmark is a total failure. My score on stock settings is 47 seconds. My second weaker computer has a score of 42 seconds. In addition, when comparing both of these computers in other benchmarks, they show that the set with AMD is much better, but the corona does not confirm it. What is the reason for this?
My computer composition is:

CPU - AMD Ryzen Threadripper 2990WX
motherboard - ASUS ROG zenith extreme alpha
GPU - GeForce RTX 2080Ti Dual 11GB DDR6
COOLING CPU - EVGA CLC 280
RAM - Corsair DDR4 128GB 2666-16 Vengeance LPX
POWER SUPPLY - Corsair HX1200 1200W

Thank you in advance for your help! :)

2019-03-20, 21:21:35
Reply #291

lolec

  • Active Users
  • **
  • Posts: 81
    • View Profile
I would recomend you do more research on RAM.

Threadripper (and it looks like specially corona) are extremely sensitive to ram frequency / # of channels.

I'm too lazy to do a full research but I can give you a summary of my experience using 128gb of Corsair Vengance memory with the 2990wx.

Although I have 3200MHz memory, I couldn't get it to work over 3000 MHz (again, to lazy to tweak and tweak)

These are Corona Benchmark results as a refference.

Single stick at 3000MHz ( 1:12)
2 sticks at 3000MHz (1:08)
4 sticks at 3000MHz (0:38)
8 sticks at 3000MHz  (Blue screen :( )
8 sticks at 2166MHZ (1.06)

As you can see, you need to populate 4 channels to reach the full potential of the processor (not a problem since you are planning to populate all of them... however, frequency also has a big impact)

I'm not sure why I couldn't make all 8 sticks to work at 3000, someone said the corsair vengance might not me the most compatible modules.

My recomendation would be to get 3200MHz momory and reasearch which brand works best to ensure you can run 8 of them at that freq.



2019-03-21, 10:27:48
Reply #292

DUMNY

  • Users
  • *
  • Posts: 3
    • View Profile
I know, I have read on the corona forum about this memory. The only thing that puzzles me is that my second computer (2x XEON E5-2696) in every benchmark is a lot slower, but when it comes to the corona amd 2990wx definitely loses. For me, the conclusion is very simple - it is not memory problem in my case, because in other tests the processor counts without major problems.
For me, corona render makes AMD 2990wx so slow, but I do not know what the exact reason is. Do you have any other explanation for this phenomenon? Why exactly corona causes such a problem?

2019-03-21, 12:41:06
Reply #293

agentdark45

  • Active Users
  • **
  • Posts: 457
    • View Profile
I know, I have read on the corona forum about this memory. The only thing that puzzles me is that my second computer (2x XEON E5-2696) in every benchmark is a lot slower, but when it comes to the corona amd 2990wx definitely loses. For me, the conclusion is very simple - it is not memory problem in my case, because in other tests the processor counts without major problems.
For me, corona render makes AMD 2990wx so slow, but I do not know what the exact reason is. Do you have any other explanation for this phenomenon? Why exactly corona causes such a problem?

The Zen architecture is heavily dependent on memory speed and latency. This is even more crucial with Threadripper due to the infinity fabric link of the CCX core modules. The issue you are having is because the IMC of the 2990wx cannot handle 128gb of RAM at high speeds. The best RAM speed you are likely to get with 128gb is 2933mhz @ CL14 latency, however looking at the RAM you've listed you're aren't going to get close to these figures.

Intel chips are monolithic (i.e no links between separate core modules), therefore RAM speeds do not affect the chips as much.
« Last Edit: 2019-03-21, 12:45:17 by agentdark45 »
Vray who?

2019-03-21, 14:43:32
Reply #294

Gr100

  • Active Users
  • **
  • Posts: 83
    • View Profile
Hello, all corona freaks

For several weeks I have been working on a new machine that I have made myself. The rendering engines I mainly use are v-ray and of course corona. Vray benchmark definitely satisfies me, unfortunately the corona benchmark is a total failure. My score on stock settings is 47 seconds. My second weaker computer has a score of 42 seconds. In addition, when comparing both of these computers in other benchmarks, they show that the set with AMD is much better, but the corona does not confirm it. What is the reason for this?
My computer composition is:

CPU - AMD Ryzen Threadripper 2990WX
motherboard - ASUS ROG zenith extreme alpha
GPU - GeForce RTX 2080Ti Dual 11GB DDR6
COOLING CPU - EVGA CLC 280
RAM - Corsair DDR4 128GB 2666-16 Vengeance LPX
POWER SUPPLY - Corsair HX1200 1200W

Thank you in advance for your help! :)

Thats crazy, im running the same system here pretty much. Discouraging after paying $6300 CAD for it. On the flip side, im able to render out a pretty heavy exterior scene at 4K in about 45 mins with about 2% noise level. About to run the bechmark to see what i get.

2019-03-21, 16:34:04
Reply #295

tallbox

  • Active Users
  • **
  • Posts: 96
  • George Nicola
    • View Profile
    • Architectural Visualisations TALLBOX
Hello, all corona freaks

For several weeks I have been working on a new machine that I have made myself. The rendering engines I mainly use are v-ray and of course corona. Vray benchmark definitely satisfies me, unfortunately the corona benchmark is a total failure. My score on stock settings is 47 seconds. My second weaker computer has a score of 42 seconds. In addition, when comparing both of these computers in other benchmarks, they show that the set with AMD is much better, but the corona does not confirm it. What is the reason for this?
My computer composition is:

CPU - AMD Ryzen Threadripper 2990WX
motherboard - ASUS ROG zenith extreme alpha
GPU - GeForce RTX 2080Ti Dual 11GB DDR6
COOLING CPU - EVGA CLC 280
RAM - Corsair DDR4 128GB 2666-16 Vengeance LPX
POWER SUPPLY - Corsair HX1200 1200W

Thank you in advance for your help! :)

Thats crazy, im running the same system here pretty much. Discouraging after paying $6300 CAD for it. On the flip side, im able to render out a pretty heavy exterior scene at 4K in about 45 mins with about 2% noise level. About to run the bechmark to see what i get.
I have ordered a very similar build a few days ago and tomorrow is arriving. I've checked corona benchmark for 2990wx and it gets decent results, where do you see problems with the speed? P
PS: Thanks to the forum I purchased 2933 mhz memory.
Cheers guys
Architectural Story Telling / Deep work practitioner
https://www.tallbox.co.uk

2019-03-21, 19:18:46
Reply #296

Eugene Danin

  • Users
  • *
  • Posts: 4
    • View Profile

2019-03-22, 02:08:01
Reply #297

DUMNY

  • Users
  • *
  • Posts: 3
    • View Profile
I'll show you what I mean the last time. Because I have the impression that either I do not understand you or you me: P. Look at the results of my two computers - 3 simple tests, I can do more if you want.

AMD 2990WX

cinebench 15R - 4900-5000 score
V-ray benchmark - 27 sec.
Corona render - 45 sec - 60 sec (here completely do these times differ from each other)

2x XEON E5-2696
cinebench 15R - 4250-4300 score
V-ray benchmark - 35 sec.
Corona render - 43 sec - 45 sec

You say that I have too slow RAM, but what does it refer to this comparison? In all tests, I should have a problem, not only with corona renderer. Maybe it's a matter of windows 10, or maybe some other setting that affects counting processes in the corona. The worst thing about all this is that I use this engine most often in recent times.:/

2019-03-22, 12:30:54
Reply #298

tallbox

  • Active Users
  • **
  • Posts: 96
  • George Nicola
    • View Profile
    • Architectural Visualisations TALLBOX
I'll show you what I mean the last time. Because I have the impression that either I do not understand you or you me: P. Look at the results of my two computers - 3 simple tests, I can do more if you want.

AMD 2990WX

cinebench 15R - 4900-5000 score
V-ray benchmark - 27 sec.
Corona render - 45 sec - 60 sec (here completely do these times differ from each other)

2x XEON E5-2696
cinebench 15R - 4250-4300 score
V-ray benchmark - 35 sec.
Corona render - 43 sec - 45 sec

You say that I have too slow RAM, but what does it refer to this comparison? In all tests, I should have a problem, not only with corona renderer. Maybe it's a matter of windows 10, or maybe some other setting that affects counting processes in the corona. The worst thing about all this is that I use this engine most often in recent times.:/

Hi, have you checked this if it's working for you: https://www.bit-tech.net/news/tech/cpus/coreprio-tool-near-doubles-threadripper-2990wx-performance/1/
Architectural Story Telling / Deep work practitioner
https://www.tallbox.co.uk

2019-03-22, 16:18:55
Reply #299

agentdark45

  • Active Users
  • **
  • Posts: 457
    • View Profile
I'll show you what I mean the last time. Because I have the impression that either I do not understand you or you me: P. Look at the results of my two computers - 3 simple tests, I can do more if you want.

AMD 2990WX

cinebench 15R - 4900-5000 score
V-ray benchmark - 27 sec.
Corona render - 45 sec - 60 sec (here completely do these times differ from each other)

2x XEON E5-2696
cinebench 15R - 4250-4300 score
V-ray benchmark - 35 sec.
Corona render - 43 sec - 45 sec

You say that I have too slow RAM, but what does it refer to this comparison? In all tests, I should have a problem, not only with corona renderer. Maybe it's a matter of windows 10, or maybe some other setting that affects counting processes in the corona. The worst thing about all this is that I use this engine most often in recent times.:/

Hi, have you checked this if it's working for you: https://www.bit-tech.net/news/tech/cpus/coreprio-tool-near-doubles-threadripper-2990wx-performance/1/

An interesting test would be to see how to 2950x stacks up against the 2990wx at the same frequency. This would indicate if there is indeed a bottleneck in corona with the 2990wx's memory sharing configuration.
Vray who?