Intel Skylake / Kaby Lake

[coercitiv]

I would like to know temperatures when the CPU is limited to 140W its rated TDP.

I would like to know both temps & performance when a hard 140W TDP limit is in place.

 

[USER8000]

So Ryzen is a multipurpose CPU, that wins in many productivity benchmarks, looses in games by a small margin and uses 66% of the power. Oh, and costs less too.

You can spin it any way you want.

It also sadly for him also uses a sub 200MM2 chip as opposed to one which is over 320MM2,so this is why AMD can get away with pricing it cheaper whilst only using dual channel RAM. Since Intel is using a larger chip,and a better process,I would certainly expect them to win since they are throwing more transistors at the problem,more die area and far more memory bandwidth.

 

[lobz]

Awesome. He was able to get a VRMs to overheat in an incredibly unrealistic usage scenario. Guess what? I can blow up the greatest of car engines if you remove the rev limiter.

Awesome, these big words, yet your kind of user is the one that if they purchased a delidded&OC'd product, they would immediately torture it the cruelest way possible to see if everything checks out, am I right? Where does he make the mistake exactly?

Sent from my VTR-L09 using Tapatalk

 

[SpoCk0nd0pe]

Awesome. He was able to get a VRMs to overheat in an incredibly unrealistic usage scenario. Guess what? I can blow up the greatest of car engines if you remove the rev limiter.

Not at all. Heavily optimized compute tasks can sometimes get even more out of the chip's ALUs.

The correct analogy is a car advertised as incredibly fast but the engine blowing out at 200 kmh. If you refer to gaming loads then it is like noone noticing because the car is driven by your grandma in an inner city traffic jam.

 

[Eug]

I'm not in the X class. Much lower end for me, but in an all-in-one iMac (as mentioned earlier).

I had the 7700K but with heavy load the CPU would hit 100C quickly and sit around 95C with the fan at max 2700 rpm - loud. This all happens in well under one minute. Furthermore, even with some other more bursty stuff - not video encoding - I could occasionally get the fan to ramp up somewhat.

So I returned it and bought a 7600 non-K. Now with heavy load the CPU maxes out initially at around 68C and then slowly climbs from there. After pegging the CPU for 10 minutes with software HEVC encoding, the temp had climbed to the low 80s, but it was plateauing out, and the fan was still at the minimum, 1200 rpm and silent. I'm not sure if it would have ever reached 90C, and certainly, the fan would never hit max. (The fan in the iMac 7600 is apparently the same as the fan in the iMac 7700K.)

Sure, stuff like encoding takes longer on the 7600, but overall I'm happier with the slower machine.

It would also be nice to see a 7700T in a Mac mini at some point.

Also, my Core m3-7Y32 has been serving me well. It's great being able to play high bitrate 10-bit 4K 2160p60 HEVC video with low CPU usage in a fanless machine.

 

[TheF34RChannel]

OC3D together with Der8auer:

 

[Zucker2k]

OC3D together with Der8auer:

It looks like this guy was essentially happily priming at 4.5ghz without even bothering with the AVX offset.

PS: OFF TOPIC: I've always been skeptical of Ryzen's treatment of Prime 95 (because of the power consumption figures, and The Stilt even mentioned it in the 'Ryzen Technical' thread ), and hearing him say this got me wondering whether AMD may have programmed Ryzen to run Prime 95 with a negative offset. I believe running the benchmark portion of Prime 95 should shed some better light on this matter.

 

[FIVR]

Conspiracy theories? LOL


"AMD programmed Ryzen to secretly run at lower temperatures to make intel look bad when they launched skylake-X" is probably the funniest thing I've read on these tech forums.


Skylake X is pretty much a disaster at this point, no conspiracy theories required to see why Ryzen is beating it in so many perf/w tests. Heck Intel can't even beat BW-E in perf/w... maybe even worse than Haswell. What's next, a 220W chip?

 

[mikk]

PS: OFF TOPIC: I've always been skeptical of Ryzen's treatment of Prime 95 (because of the power consumption figures, and The Stilt even mentioned it in the 'Ryzen Technical' thread ), and hearing him say this got me wondering whether AMD may have programmed Ryzen to run Prime 95 with a negative offset. I believe running the benchmark portion of Prime 95 should shed some better light on this matter.

They don't need that because Ryzen doesn't run AVX1/AVX2 full speed. So they don't have these heat issues there but they also suffer from a poor performance if AVX2 is properly used, recent x265 binaries for example.

 

[DrMrLordX]

hearing him say this got me wondering whether AMD may have programmed Ryzen to run Prime 95 with a negative offset. I believe running the benchmark portion of Prime 95 should shed some better light on this matter.

To hell with Prime95, anyone with a Ryzen should be using the latest version of y-cruncher for heat/stability testing. I am somewhat annoyed that we don't see more Skylake-X testers using it. You can select which ISA extension you want as well by selecting the correct binary. Want Skylake-X to run SSE4.1? Go for it!

 

[AdamK47]

OC3D together with Der8auer:

I'm surprised he didn't know to disable SVID and to bump up the CPU current percentage. These are things I always set after every BIOS upgrade since first using X99 in late '14. I don't go to extremes on my overclocks (6950X 4.125GHz @ 1.225V). I do always test my overclocks with Prime95 small FFT. Sometimes I even go beyond that and free up a couple cores and loop 3DMark while running Prime95. Going into worst case scenarios is what I've always done to test overclocks.

 

[Arachnotronic]

It looks like this guy was essentially happily priming at 4.5ghz without even bothering with the AVX offset.

Honestly, I'm impressed that these CPUs can even handle AVX loads at those frequencies. They eat up power/generate lots of heat...but that IS a lot of SIMD compute power.

 

[AdamK47]

One other thing I liked about that video is his rant against Asus. Some of the Asus representatives (ahem... Raja) can be completely obstinate when it comes to testing overclocks. They don't believe in worst case scenarios like running Prime95, LinX, OCCT, or anything with linpack AVX. They push as much voltage as they can to achieve higher overclocks and run weak tests like Realbench (of course), AIDA64, or multiple instances of HCI Memtest. I'm willing to bet these Asus guys are right now going against OC3D and der8auer's claims.

 

[Zucker2k]

Conspiracy theories? LOL


"AMD programmed Ryzen to secretly run at lower temperatures to make intel look bad when they launched skylake-X" is probably the funniest thing I've read on these tech forums.


Skylake X is pretty much a disaster at this point, no conspiracy theories required to see why Ryzen is beating it in so many perf/w tests. Heck Intel can't even beat BW-E in perf/w... maybe even worse than Haswell. What's next, a 220W chip?

It's interesting how you misinterpreted my comments and put them in a nice quote. This is quite fraudulent, hehe.

P95 has always been, by far, the holy grail of cpu stress-testing since as far back as I can remember, and continues to be. So, of course, I was surprised to see Ryzen pulling significantly less watts than it should, running the software; and I don't think it is limited to the avx version alone either. As has been pointed out, though, if Ryzen is only running avx code at half-speed, then how can anyone reasonably expect to see anything other than the numbers we're seeing from SKL-X? P95 torture test is a power virus running the thirstiest code in P95 (avx version) at insane overclocks.

You know what? I'll concede if you can show me a Ryzen chip on the same cooling as the 7820x in Tahoe's sig, clocked anywhere above 4.3ghz running P95 without throttling. Now, that's mission impossible right there!

 

[.vodka]

You know P95 wasn't updated to support Ryzen until very recently. P95 testing with Ryzen was done at launch, with it showing abnormally low power consumption for a stress test.

29.2 (June 10th, 2017)

• Throughput benchmarks of all FFT implementations are written to gwnum.txt.
• These benchmarks will then be used to pick the best FFT implementation to use.
• This can lead to minor performance improvements.
• Benchmarks are run daily for any FFT sizes that will be needed within the next 7 days.
• Once enough benchmark data is accumulated, automatic benchmarks no longer take place.
• Default FFT implementations for Ryzen added.
• A few new FFT sizes (up to 50M) are enabled for FMA3-capable CPUs.

29.1 (March 27th, 2017)

• Faster trial factoring for machines that support FMA (Haswell and later). Multi-threaded trial factoring now supports more than one thread sieving for small primes. Several tuning parameters added - see undoc.txt.
• The portable library, hwloc, for analyzing a machine's topology is now used. This replaces the buggy code prime95 used to detect hyperthreading. It also eliminates the need for AffinityScramble2. Running a benchmark will output this topology information to results.txt.
• AVX-512 trial factoring support added.
• Dialog box for benchmarking added.
• In the Test/Worker Windows dialog box you no longer choose how many threads each worker uses. Instead, you choose how many CPU cores each worker uses. There affinity options have been removed. There are two new options that will decide if each worker also uses hyperthreading.

Someone with Ryzen should test 29.2 and see how much it drives power consumption up. Same for those with SKL-X as they get a AVX512 workout... especially the 7900x without execution penalty vs the 6c and 8c parts.

der8auer's P95 power numbers were done with 27.1... this version implements up to Sandy/Ivy AVX... the AVX2/FMA/AVX512 hardware in SKL-X goes completely unutilized. 4.5GHz 1.25v 7900x under those conditions and running small FFTs measured 502w at the wall . I don't even want to know how much those power numbers would shoot up with these parts of the CPU running full tilt and having the CPU configured to not throttle.

 

[Markfw]

Well, I don't have my kill-a-watt on it right now, but its running and 52c is the highest its gotten.

The Ryzen 1800x

Edit, its now hit 55c after 16 minutes

 

[Markfw]

OK, I shut it down and put the kill-a-watt in. Idle, 65 watts, full load 170 watts. Temps went from 40c to 55c

Edit: This is the complete system, from the wall, and yes thats all 8 cores and 16 threads.

Oh, this is running blend. Do you want something else ? Task manager says its running 100% CPU

 

[TahoeDust]

OK, I shut it down and put the kill-a-watt in. Idle, 65 watts, full load 170 watts. Temps went from 40c to 55c

Edit: This is the complete system, from the wall, and yes thats all 8 cores and 16 threads.

Oh, this is running blend. Do you want something else ? Task manager says its running 100% CPU

What speed is it actually running AVX at?

 

[Markfw]

What speed is it actually running AVX at?

No idea, just running 29.2 Prime95

 

[FIVR]

It's interesting how you misinterpreted my comments and put them in a nice quote. This is quite fraudulent, hehe.

P95 has always been, by far, the holy grail of cpu stress-testing since as far back as I can remember, and continues to be. So, of course, I was surprised to see Ryzen pulling significantly less watts than it should, running the software; and I don't think it is limited to the avx version alone either. As has been pointed out, though, if Ryzen is only running avx code at half-speed, then how can anyone reasonably expect to see anything other than the numbers we're seeing from SKL-X? P95 torture test is a power virus running the thirstiest code in P95 (avx version) at insane overclocks.

You know what? I'll concede if you can show me a Ryzen chip on the same cooling as the 7820x in Tahoe's sig, clocked anywhere above 4.3ghz running P95 without throttling. Now, that's mission impossible right there!

You know what? If you can find a 7820x that runs Prime95 at stable 4.3Ghz without drawing way over the rated power for Skylake-X motherboards (165W) I'd be happy to to concede that Skylake might just be a "huge mistake" and not an absolute unmitigated disaster. Too bad there isn't much chance of that

 

[VirtualLarry]

(Off-topic: Ryzen info)
I've been running PrimeGrid SG tasks (OK, not exactly the same as Prime95, I don't think). Temps are 10C higher than running WCG tasks. I have a (few) 1600 CPUs. I decided to remove the 3.8Ghz OC I had on my 1600 CPU that was under 120mm AIO liquid cooling, as it was hitting around / above 80C running PrimeGrid tasks.

 

[DrMrLordX]

I'd be more than happy to update Prime95 and run some other AVX/"ADX" benchmarks (see: y-cruncher) on my 1800x but this is a Skylake thread, so can we start a new topic on that?

 

[TahoeDust]

You know what? If you can find a 7820x that runs Prime95 at stable 4.3Ghz without drawing way over the rated power for Skylake-X motherboards (165W) I'd be happy to to concede that Skylake might just be a "huge mistake" and not an absolute unmitigated disaster. Too bad there isn't much chance of that

I can show you one that will outrun anything AMD makes, even when overclocked, while drawing 165w.

 

[Topweasel]

They don't need that because Ryzen doesn't run AVX1/AVX2 full speed. So they don't have these heat issues there but they also suffer from a poor performance if AVX2 is properly used, recent x265 binaries for example.

No offense but that is all crap. The difference between Ryzen, a 7700, and let's say a 7900x on AVX is that AMD A.) Doesn't use 512 like the 7900 does. B.) Uses 2 128bit piplines for handling 1 AVX 2 instructions.

It's not running it at separate speeds. It's just putting less stress on the system because it's not using a 256bit or 512bit system (which for the later is good, just the register file on SLx is the size of an atom processor). It's more in line with the rest of the CPU and is less likely to hot spot the CPU like it is on Intel's solution. Compared to a 7700 clock for clock a 1700 would be just as fast in AVX2 and twice as fast in AVX.

In SL-X with it catching up and exceeding the R7 in cores. SL-X is going to be a lot better. But on top of high power usage and possible cooling issues, this superiority has a cost specially when over clocking.

 

[Zucker2k]

You know what? If you can find a 7820x that runs Prime95 at stable 4.3Ghz without drawing way over the rated power for Skylake-X motherboards (165W) I'd be happy to to concede that Skylake might just be a "huge mistake" and not an absolute unmitigated disaster. Too bad there isn't much chance of that

You mean the 165w (minimum) power specification? Pffft! It's like crying over the fact that your 200mph stock corvette is actually capable of speeds of 300mph but uses more fuel to do it! What kinda logic is that? HEDT boards are built way beyond spec to accommodate extreme overclocking. They can deliver the necessary power if the chip needs it, which is what you get when you run a torture test like P95.