The cable temp is half the problem. VRM's are the other half. In any case, a single 8 pin is pushing it for 300w.
According to jonnyguru, if der8auer used the SuperFlower modular PSU cables he showed in the video, it might just be terrible cable design causing the 12V rail to drop too far under load.
PSU problem causing this uproar would be hilarious.
The cable temp is half the problem. VRM's are the other half. In any case, a single 8 pin is pushing it for 300w.
According to jonnyguru, if der8auer used the SuperFlower modular PSU cables he showed in the video, it might just be terrible cable design causing the 12V rail to drop too far under load.
PSU problem causing this uproar would be hilarious.
Wouldn't explain other people using different PSUs having the same VRM heat issues though. Could explain away the the cable heat portion for sure. It seems the EPS12V @~28 amps 336 watt limit might be a bit dicey with a 10 core overclocking past 4.5.
The other guy is Silicon Lottery right? What PSU are they using?
Yeah, definitely want EPS12V + ATX12V for the 10core and above and then set a board limit of 450W or something. Barring that set a short term limit of 320W and long term of 280W. Most of these boards should be able to adjust long term and short term turbo limits.
Good information, thank you. Is power consumption the overall system power consumption?
Note that the CPU clock and NB Clock are being incorrectly reported. They are actually 4Ghz and 2.4Ghz respectively. But the numbers are not that bad otherwise.
The cable temp is half the problem. VRM's are the other half. In any case, a single 8 pin is pushing it for 300w.
*Update* Found where I can change the NB clock (Now called CLR in BIOS). Re-ran the test with the NB set to 3200mhz. Big improvement.
Ok, had only a few hours tonight to break in my new rig, but here are my initial impressions and test results.
Stock (4Ghz all cores) - voltages set to Auto
- Very stable, and efficient. No problem OCing RAM to 3200mhz with no required bump to IMC or NB voltages.
- Ran Prime95 (FMA-AVX2) for 10 minutes using option that generates the most heat. Results are as follows:
VCore: 1.05v
Max Core Temp: 61C
Max VRM Temp: 60C
142 watts
Mild OC (4.2Ghz all cores) - voltages set to Auto
- Ran Prime95 (FMA-AVX2) for 10 minutes using option that generates the most heat. Results are as follows:
VCore: 1.10v
Max Core Temp: 67C
Max VRM Temp: 64C
151 watts
Did not have time to go any higher tonight. But to be honest, I was happy with my 6600K at 4.2Ghz (which required 1.25v to achieve), so I think I may eventually land back here as my 24/7 speed.
The Good
- Stable out of the box. Much better experience than with my z170x MB on release day two years ago. It took 3 BIOS updates later to even OC my RAM to 3000mhz. With the X299, OCing RAM to 3200 was a snap (even with an immature BIOS).
- Very energy efficient at stock and low OC speeds. Less voltage and runs cooler than my 6600K at the same speeds. I think this is a large part due to 14nm+.
- Different offsets for AVX and AVX-512. This can really help achieve better OC results in my opinion.
The Bad
- Northbrige runs at 2.4Ghz regardless of CPU speed. And unlike all past platforms, this speed is not changeable in the BIOS. Maybe this is simply a Gigabyte BIOS issue, but there is no place to change that stock value.
- You can not disable the AVX offset. You either have to choose a value from 1 to 31 (1 being 100mhz less than CPU frequency) or choose AUTO (which sets the offset to 300mhz). So you can never run AVX code at CPU speeds. Again, this may be fixed in later BIOS updates and not be a limitation of the platform itself.
- NVMe drives running slow. Same drives on a Z170 MB were running much faster (see my post above). Using the same test program (Aida64) and same drives, speed decreased by over 1000MBps. This may be an Aida issue or again a BIOS issue. I will try other tests tomorrow.
The Ugly
- None yet. But other people's reviews of the 7900X is concerning.
Thoughts
So far I am happy. I expected issues as an early adopter so I am not concerned with the few issues I have found. I am also not experiencing the same heat/power/vrm issues with the 7820X as others are having with the 7900X. I will do more testing tomorrow after work.
Thanks for the write up. I was hoping that the uncore/L3 speed would be overclockable but I suppose not! My parts get in today, hopefully I'll have a smooth time getting it all set up.
Read my update....it is. And by over clocking it made a big difference in the results.
That is a nice improvement. Any chance you could run Intel memory latency checker on it?
https://software.intel.com/en-us/articles/intelr-memory-latency-checker
I am esp interested in "loaded" latencies.
EDIT: it is best to run it from administrative command prompt and non avx512 version i guess?
EDIT2: I would be VERY grateful if You could also run and report the results on the following commands:
mlc --c2c_latency -c0 -w1
mlc --c2c_latency -c0 -w2
mlc --c2c_latency -H -c0 -w1
mlc --c2c_latency -H -c0 -w2
When I get home tonight I will try to run these for you.
Will likely be evening time.
I am having trouble with some of these X299 motherboards. I've bought a wide variety for this launch, and none of them are really handling the load of an overclocked 7900X as well as I'd expect. VRM temps through the roof and boards throttling.
So far so good.
Stock (4Ghz all cores) and running Prime95 (full FMA3 mode) and my temps are not going over 60C (pulling 144 watts). I am having trouble with the voltage readings as cpuz and aids64 both are giving me incorrect readings.
Got RAM to 3200mhz without issue (no voltage bumps required on the IMC).
Going to try OCing soon, once I get a reliable reading on voltage.
May I ask if you had to raise your BCLK to get the RAM to run at 3200? And did you use an XMP profile or just manually OC it?
And also, isn't all-core turbo 4.3 GHz stock? I was confused when you said you OC'd it to 4.2. Got my 7820 on order and trying not to get too scared by all the negativity. Thanks!
I wonder why the default NB (Mesh) clock is 2400mhz. Seems pretty low considering the NB on Skylake-S was defaulted to 3900mhz (yes I know it was a ring bus).
Since there is zero data on Skylake-X mesh overclocks, I will do more experimenting tonight.
Single core turbo is 4.3 at stock. All core turbo is 4.0 at stock.
I did not use XMP to get my RAM to 3200. I did not have to change any BCLK settings. I simply changed my clock speed and DRAM voltage to 1.35v.
Don't be scared. At stock or mild OCs, this is a very stable platform for me so far.
I'd have thought it would at least track the RAM data rate, since if interconnect frequency < DRAM rate it would bottleneck memory access.
How was the power when overclocking mesh?
No voltage changes required. I did not run a stress test on it yet with the OC mesh (I literally changed it before I had to run to work this morning because it was driving me crazy all night). I will post more results tonight with the mesh @ 3200mhz.
Thanks! All the retailers are advertising 4.3 on all cores, 4.5 on up to 2 cores - I must be misunderstanding something there. Are you at least seeing Turbo Boost 3.0 work for 4.5 GHz on 2 cores?
Also what PSU are you using? I was looking at the EVGA SuperNOVA 850 or 1000 G3 however they are made by SuperFlower which scares me because der8auer was apparently having his temperature issues with a SuperFlower unit, but I am not sure if it was an EVGA or exactly what unit it was.
Also did you get your 960 Pro running at max speed?
Thanks for the help, sorry for all the questions!