Intel Core X TDP 165W and noise

[Caffreys]

Hello,

I would like to know ... the owners of Core i9-99?0X processors with a TDP of 165W,
did Prime95 run on all the cores without the noise being unbearable ?

Liquid cooling? which model ?
How did you realize that ?

Currently I have a Xeon E5-1660v2 with TDP 130W but it's ok.

I would like to buy a Core i9-10980X 18 cores but I'm afraid of noise with the computer just beside me.

Thank you !

[Mysticial]

If you're not overclocking it, you can pop any large AIO on it and it'll stay quiet.

And if you intend to run LL on it, it'll be memory-bound anyways. So it won't be able to sustain a high power draw.

[Caffreys]

Originally Posted by Mysticial

So it won't be able to sustain a high power draw.

But DDR4-2933 quad channel is probably enough for 18 cores at 3Ghz ?! no ?
You mean that the cores will not be close to 100%-97% use ?

[Mysticial]

Originally Posted by Caffreys

But DDR4-2933 quad channel is probably enough for 18 cores at 3Ghz ?! no ?
You mean that the cores will not be close to 100%-97% use ?

Not even close. You'd probably need 8 - 16 channels to keep LL from bottlenecking on 18-core Skylake X.

The high core count Skylake X's aren't a good value for LL for this reason. You're probably better off with perhaps a 10-core and a maxed out memory overclock like 3600+ MT/s.

[Caffreys]

Ok Thanks


In any case, the most powerful processor Skylake X does not exceed twice my Xeon E5-1660 v2 Ivy-Bridge;
I will wait.

[scan80269]

How much throughput (ms/iter) for Prime95 are you getting with your Xeon E5-1660 v2?

Since this Xeon is paired with DDR3 memory, and Prime95 at the wavefront (FFT=5120K) is memory bandwidth bound with Intel HEDT systems, I believe a system using 10-core Skylake-X Core i9-9900X or the imminent Cascade Lake-X Core i9-10900X paired with quad-channel DDR4 memory at 2666/2933 (or higher with overclocking) and utilizing AVX-512 should be able to achieve at least 2X throughput vs. this Xeon.

I am currently very tempted to go for i9-10900X myself, and pair it with DDR4-3600 or 3733 dual-ranked memory. If everything works out and the system is stable, I should have some throughput numbers for Cascade Lake-X to report in a while.

[Caffreys]

On my Xeon E5-1660 v2 16GB DDR3-1866 CL13

Prime95 v29.4 build 8
I get

Timings for 5120K FFT length (6 cores, 1 worker):  6.81 ms.  Throughput: 146.80 iter/sec.
Timings for 5120K FFT length (6 cores, 6 workers): 39.46, 39.26, 39.17, 39.14, 39.26, 39.25 ms.  Throughput: 152.85 iter/sec.
Timings for 5376K FFT length (6 cores, 1 worker):  6.96 ms.  Throughput: 143.58 iter/sec.
Timings for 5376K FFT length (6 cores, 6 workers): 41.10, 41.08, 41.06, 40.85, 41.07, 41.08 ms.  Throughput: 146.20 iter/sec.
Timings for 5760K FFT length (6 cores, 1 worker):  7.22 ms.  Throughput: 138.49 iter/sec.
Timings for 5760K FFT length (6 cores, 6 workers): 42.77, 42.64, 42.71, 42.59, 42.61, 42.61 ms.  Throughput: 140.67 iter/sec.
Timings for 6144K FFT length (6 cores, 1 worker):  8.22 ms.  Throughput: 121.71 iter/sec.
Timings for 6144K FFT length (6 cores, 6 workers): 47.43, 47.71, 47.40, 47.41, 47.39, 47.38 ms.  Throughput: 126.44 iter/sec.
Timings for 6400K FFT length (6 cores, 1 worker):  8.21 ms.  Throughput: 121.79 iter/sec.
Timings for 6400K FFT length (6 cores, 6 workers): 48.78, 48.47, 48.53, 48.63, 48.49, 48.47 ms.  Throughput: 123.55 iter/sec.
Timings for 6720K FFT length (6 cores, 1 worker):  9.08 ms.  Throughput: 110.16 iter/sec.
Timings for 6720K FFT length (6 cores, 6 workers): 50.46, 50.22, 50.29, 50.51, 50.23, 50.22 ms.  Throughput: 119.23 iter/sec.
Timings for 6912K FFT length (6 cores, 1 worker):  9.47 ms.  Throughput: 105.65 iter/sec.
Timings for 6912K FFT length (6 cores, 6 workers): 60.51, 60.55, 60.52, 60.44, 60.56, 60.41 ms.  Throughput: 99.17 iter/sec.
Timings for 7168K FFT length (6 cores, 1 worker):  9.67 ms.  Throughput: 103.45 iter/sec.
Timings for 7168K FFT length (6 cores, 6 workers): 56.03, 55.93, 55.97, 55.71, 55.91, 55.84 ms.  Throughput: 107.34 iter/sec.
Timings for 7680K FFT length (6 cores, 1 worker):  9.88 ms.  Throughput: 101.21 iter/sec.
Timings for 7680K FFT length (6 cores, 6 workers): 58.60, 58.34, 58.41, 58.28, 58.18, 58.20 ms.  Throughput: 102.85 iter/sec.
Timings for 8000K FFT length (6 cores, 1 worker): 10.28 ms.  Throughput: 97.26 iter/sec.
Timings for 8000K FFT length (6 cores, 6 workers): 60.67, 60.15, 59.90, 59.90, 59.90, 60.36 ms.  Throughput: 99.76 iter/sec.
Timings for 8192K FFT length (6 cores, 1 worker): 11.29 ms.  Throughput: 88.55 iter/sec.
Timings for 8192K FFT length (6 cores, 6 workers): 65.59, 65.03, 65.35, 65.25, 64.88, 65.10 ms.  Throughput: 92.03 iter/sec.

[fivemack]

On my i9-7940X v29.8 build 6

14 cores

[Worker #1 Nov 5 20:05] Timing 25 iterations of 5120K FFT length.  Best time: 2.264 ms., avg time: 2.428 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 5184K FFT length.  Best time: 2.430 ms., avg time: 2.611 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 5376K FFT length.  Best time: 2.458 ms., avg time: 2.618 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 5760K FFT length.  Best time: 3.308 ms., avg time: 3.422 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 6048K FFT length.  Best time: 3.051 ms., avg time: 3.149 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 6144K FFT length.  Best time: 3.122 ms., avg time: 3.233 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 6272K FFT length.  Best time: 3.198 ms., avg time: 3.288 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 6400K FFT length.  Best time: 3.545 ms., avg time: 3.618 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 6720K FFT length.  Best time: 3.757 ms., avg time: 3.827 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 7056K FFT length.  Best time: 3.995 ms., avg time: 4.049 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 7168K FFT length.  Best time: 3.996 ms., avg time: 4.055 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 7200K FFT length.  Best time: 3.930 ms., avg time: 3.985 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 7680K FFT length.  Best time: 4.352 ms., avg time: 4.400 ms.
[Worker #1 Nov 5 20:05] Timing 25 iterations of 8064K FFT length.  Best time: 5.064 ms., avg time: 5.122 ms.

7 cores

[Worker #1 Nov 5 20:06] Timing 25 iterations of 5120K FFT length.  Best time: 2.755 ms., avg time: 3.041 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 5184K FFT length.  Best time: 2.996 ms., avg time: 3.247 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 5376K FFT length.  Best time: 2.923 ms., avg time: 3.139 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 5760K FFT length.  Best time: 3.911 ms., avg time: 4.110 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 6048K FFT length.  Best time: 3.428 ms., avg time: 3.609 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 6144K FFT length.  Best time: 3.530 ms., avg time: 3.778 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 6272K FFT length.  Best time: 3.587 ms., avg time: 3.857 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 6400K FFT length.  Best time: 3.899 ms., avg time: 4.051 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 6720K FFT length.  Best time: 4.099 ms., avg time: 4.303 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 7056K FFT length.  Best time: 4.294 ms., avg time: 4.521 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 7168K FFT length.  Best time: 4.290 ms., avg time: 4.423 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 7200K FFT length.  Best time: 4.239 ms., avg time: 4.369 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 7680K FFT length.  Best time: 4.649 ms., avg time: 4.829 ms.
[Worker #1 Nov 5 20:06] Timing 25 iterations of 8064K FFT length.  Best time: 5.334 ms., avg time: 5.495 ms.

2 cores

[Worker #1 Nov 5 20:07] Timing 25 iterations of 5120K FFT length.  Best time: 7.925 ms., avg time: 8.047 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 5184K FFT length.  Best time: 8.665 ms., avg time: 8.795 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 5376K FFT length.  Best time: 8.333 ms., avg time: 8.461 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 5760K FFT length.  Best time: 10.930 ms., avg time: 11.065 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6048K FFT length.  Best time: 9.617 ms., avg time: 9.722 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6144K FFT length.  Best time: 9.649 ms., avg time: 9.755 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6272K FFT length.  Best time: 9.852 ms., avg time: 9.981 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6400K FFT length.  Best time: 10.337 ms., avg time: 10.458 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6720K FFT length.  Best time: 10.916 ms., avg time: 11.003 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 7056K FFT length.  Best time: 11.681 ms., avg time: 11.786 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 7168K FFT length.  Best time: 11.404 ms., avg time: 11.501 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 7200K FFT length.  Best time: 11.503 ms., avg time: 11.572 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 7680K FFT length.  Best time: 12.227 ms., avg time: 12.312 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 8064K FFT length.  Best time: 14.015 ms., avg time: 14.114 ms.

4 cores

[Worker #1 Nov 5 20:07] Timing 25 iterations of 5120K FFT length.  Best time: 4.291 ms., avg time: 4.472 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 5184K FFT length.  Best time: 4.669 ms., avg time: 4.883 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 5376K FFT length.  Best time: 4.483 ms., avg time: 4.645 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 5760K FFT length.  Best time: 5.911 ms., avg time: 6.080 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6048K FFT length.  Best time: 5.124 ms., avg time: 5.315 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6144K FFT length.  Best time: 5.162 ms., avg time: 5.383 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6272K FFT length.  Best time: 5.277 ms., avg time: 5.456 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6400K FFT length.  Best time: 5.560 ms., avg time: 5.757 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 6720K FFT length.  Best time: 5.853 ms., avg time: 6.035 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 7056K FFT length.  Best time: 6.266 ms., avg time: 6.474 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 7168K FFT length.  Best time: 6.056 ms., avg time: 6.247 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 7200K FFT length.  Best time: 6.083 ms., avg time: 6.260 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 7680K FFT length.  Best time: 6.549 ms., avg time: 6.698 ms.
[Worker #1 Nov 5 20:07] Timing 25 iterations of 8064K FFT length.  Best time: 7.509 ms., avg time: 7.652 ms.

[scan80269]

On my i7-5960X (Haswell-E) with DDR4-2133 CL13 4x8GB 2R memory, Prime95 v29.8 build 6:

Timings for 5120K FFT length (8 cores, 1 worker):  3.80 ms.  Throughput: 262.85 iter/sec.
Timings for 5120K FFT length (8 cores, 8 workers): 31.83, 31.70, 31.70, 31.67, 31.74, 31.87, 31.59, 31.71 ms.  Throughput: 252.16 iter/sec.
Timings for 5376K FFT length (8 cores, 1 worker):  4.18 ms.  Throughput: 239.16 iter/sec.
Timings for 5376K FFT length (8 cores, 8 workers): 34.27, 34.17, 33.99, 34.10, 34.02, 34.29, 34.19, 34.11 ms.  Throughput: 234.31 iter/sec.
Timings for 5760K FFT length (8 cores, 1 worker):  4.67 ms.  Throughput: 214.18 iter/sec.
Timings for 5760K FFT length (8 cores, 8 workers): 36.88, 36.72, 36.58, 36.75, 36.68, 36.97, 36.80, 36.87 ms.  Throughput: 217.51 iter/sec.
Timings for 6144K FFT length (8 cores, 1 worker):  4.91 ms.  Throughput: 203.74 iter/sec.
Timings for 6144K FFT length (8 cores, 8 workers): 39.03, 39.00, 38.89, 38.97, 39.04, 39.14, 39.08, 39.04 ms.  Throughput: 205.00 iter/sec.
Timings for 6400K FFT length (8 cores, 1 worker):  5.18 ms.  Throughput: 193.18 iter/sec.
Timings for 6400K FFT length (8 cores, 8 workers): 41.29, 41.37, 41.25, 41.35, 41.32, 41.40, 41.35, 41.28 ms.  Throughput: 193.58 iter/sec.
Timings for 6720K FFT length (8 cores, 1 worker):  5.49 ms.  Throughput: 182.04 iter/sec.
Timings for 6720K FFT length (8 cores, 8 workers): 43.62, 43.81, 43.23, 43.78, 43.66, 43.64, 43.80, 43.81 ms.  Throughput: 183.20 iter/sec.
Timings for 6912K FFT length (8 cores, 1 worker):  5.59 ms.  Throughput: 178.79 iter/sec.
Timings for 6912K FFT length (8 cores, 8 workers): 44.30, 44.06, 43.93, 44.34, 44.04, 44.26, 44.06, 44.73 ms.  Throughput: 180.93 iter/sec.
Timings for 7168K FFT length (8 cores, 1 worker):  5.97 ms.  Throughput: 167.62 iter/sec.
Timings for 7168K FFT length (8 cores, 8 workers): 47.03, 46.99, 47.24, 47.09, 47.09, 47.16, 46.85, 47.10 ms.  Throughput: 169.96 iter/sec.
Timings for 7680K FFT length (8 cores, 1 worker):  5.99 ms.  Throughput: 166.93 iter/sec.
Timings for 7680K FFT length (8 cores, 8 workers): 48.17, 48.36, 48.13, 48.53, 48.02, 48.39, 47.87, 48.24 ms.  Throughput: 165.94 iter/sec.
Timings for 8064K FFT length (8 cores, 1 worker):  6.46 ms.  Throughput: 154.72 iter/sec.
Timings for 8064K FFT length (8 cores, 8 workers): 51.94, 51.58, 51.47, 51.84, 51.75, 52.13, 51.53, 52.04 ms.  Throughput: 154.49 iter/sec.
Timings for 8192K FFT length (8 cores, 1 worker):  7.47 ms.  Throughput: 133.87 iter/sec.
Timings for 8192K FFT length (8 cores, 8 workers): 57.56, 57.16, 58.28, 57.59, 57.39, 57.98, 57.55, 57.66 ms.  Throughput: 138.78 iter/sec.

@Caffreys, it looks like the Prime95 throughput of your Xeon rig is bottle-necked by DDR3 memory bandwidth. A Skylake-X or Cascade Lake-X system with quad channel DDR4 memory should yield much faster iteration times.

My i7-5960X system can do better as this X99 motherboard lacks memory overclocking support, so the quad channel DDR4 is stuck running at 2133. I'm looking into a replacement X99 motherboard to get the DDR4 running at 3000 or thereabouts, which hopefully can get the iteration time for 5120K FFT (8 core, 1 worker) down below 3ms.

[scan80269]

@fivemack, what is the DDR4 memory configuration for your i9-7940X system?

[fivemack]

Originally Posted by scan80269

@fivemack, what is the DDR4 memory configuration for your i9-7940X system?

Four 16GB Corsair Vengeance LPX, advertised as 2666MHz but dmidecode says they're running at 2133MHz.

https://www.scan.co.uk/products/64gb...-16-18-18-35-x

Now, if you want silly memory bandwidth, I have a Xeon Silver 4114 system. Much less AVX512 throughput (1x at 2500MHz versus 2x at 3700MHz), but 20 cores and 12 memory channels at 2400MHz

[Worker #1 Nov 5 21:40] Timing 25 iterations of 5120K FFT length.  Best time: 2.336 ms., avg time: 2.529 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 5184K FFT length.  Best time: 2.774 ms., avg time: 3.009 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 5376K FFT length.  Best time: 2.481 ms., avg time: 2.691 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 5760K FFT length.  Best time: 3.326 ms., avg time: 3.585 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6048K FFT length.  Best time: 2.803 ms., avg time: 3.076 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6144K FFT length.  Best time: 2.764 ms., avg time: 2.990 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6272K FFT length.  Best time: 3.029 ms., avg time: 3.217 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6400K FFT length.  Best time: 3.118 ms., avg time: 3.324 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6720K FFT length.  Best time: 3.229 ms., avg time: 3.437 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 7056K FFT length.  Best time: 3.340 ms., avg time: 3.559 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 7168K FFT length.  Best time: 3.442 ms., avg time: 3.582 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 7200K FFT length.  Best time: 3.402 ms., avg time: 3.548 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 7680K FFT length.  Best time: 3.599 ms., avg time: 3.780 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 8064K FFT length.  Best time: 3.963 ms., avg time: 4.136 ms.

10 cores

[Worker #1 Nov 5 21:40] Timing 25 iterations of 5120K FFT length.  Best time: 4.159 ms., avg time: 4.293 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 5184K FFT length.  Best time: 4.444 ms., avg time: 4.586 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 5376K FFT length.  Best time: 4.437 ms., avg time: 4.568 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 5760K FFT length.  Best time: 5.219 ms., avg time: 5.376 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6048K FFT length.  Best time: 5.036 ms., avg time: 5.131 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6144K FFT length.  Best time: 4.917 ms., avg time: 5.012 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6272K FFT length.  Best time: 5.261 ms., avg time: 5.396 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6400K FFT length.  Best time: 5.301 ms., avg time: 5.486 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 6720K FFT length.  Best time: 5.606 ms., avg time: 5.764 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 7056K FFT length.  Best time: 6.007 ms., avg time: 6.122 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 7168K FFT length.  Best time: 5.915 ms., avg time: 6.016 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 7200K FFT length.  Best time: 5.966 ms., avg time: 6.039 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 7680K FFT length.  Best time: 6.221 ms., avg time: 6.303 ms.
[Worker #1 Nov 5 21:40] Timing 25 iterations of 8064K FFT length.  Best time: 6.731 ms., avg time: 6.807 ms.