Intel Skylake / Kaby Lake

[Shivansps]

Again, considering power consuption in gaming is no so bad compared to Ryzen or BDW-E.... maybe Handbrake is using AVX-512?

 

[csbin]

Skylake-X 3D Performance


https://www.bit-tech.net/hardware/2017/06/16/intel-core-i9-7900x-and-x299-chipset-revie/6

 

[csbin]

Power Consumption

 

[Bouowmx]

I guess time to pull a "wait for BIOS updates" on 06-26 or 7980XE?

 

[PhonakV30]

However, temperatures were definitely a concern with Cinebench and Terragen pushing 100°C with our 240mm AIO liquid cooler. As a result, while stable and potentially tameable under custom water-cooling, we decided to go for 4.6GHz for benchmarking, which required a super-low 1.22V. Interestingly our Core i7-6950X ran much cooler despite using a significantly higher voltage, albeit at 4.4GHz. This could well be due to thermal paste having been used between the heatspreader and CPU core with the new Skylake-X CPUs, in which case delidding could potentially yield significant benefits given the high heat density.

source : https://www.bit-tech.net/hardware/2017/06/16/intel-core-i9-7900x-and-x299-chipset-revie/8

BTW , Everything Is Ok but Temp is concern, Intel Needs to put Solder ASAP !

 

[ManyThreads]

Do you guys think that for a system with a 7820X (OC'd to hopefully 4.7), 4 X 8GB RAM, 1 X PCI SSD, 1 X SATA SSD, 4 X HDD, Titan X (Maxwell 250W), and 3 fans and a CPU Cooler (DH15) would need more than an EVGA 850W G3? Or would you bump it to 1000? The power calculators on the web are all over the map, but I think I can get away with 850.

 

[Karnak]

Again, considering power consuption in gaming is no so bad compared to Ryzen or BDW-E.... maybe Handbrake is using AVX-512?

Gaming doesn't mean anything, at least in terms of the total power consumption. For that you'll need the CPU alone since a lot of the time the GPU is at 100% load while the cpu isn't.

 

[tamz_msc]

From the Prime95 results of bit-tech, it looks like AVX-512 is in effect. Still, AVX-512, if it is working in x.264, ultimately results in worse perf/watt.

 

[R0H1T]

Do you guys think that for a system with a 7820X (OC'd to hopefully 4.7), 4 X 8GB RAM, 1 X PCI SSD, 1 X SATA SSD, 4 X HDD, Titan X (Maxwell 250W), and 3 fans and a CPU Cooler (DH15) would need more than an EVGA 850W G3? Or would you bump it to 1000? The power calculators on the web are all over the map, but I think I can get away with 850.

Too many assumptions, the 4.7GHz number is highly optimistic even with water cooling. Then there's temp & power consumption, you'll run into one or the other with that OC, chances are the D15 won't be enough though it also depends on the ambient temp.

 

[raghu78]

7900x@4.6 Ghz losing to 6950x@4.4 Ghz in half the tests is not a good sign for Skylake-X IPC. Skylake-X / Skylake-SP seems to have been designed for server workloads.

 

[Zucker2k]

No it's not ~ (112.7/76.3)*113=166.9

Though that's the platform power, it still shows how inefficient SKL-X is compared to the regular Ryzen, TR might be a different story but I doubt Intel can just make up for the difference of ~50 watts.

Yes, it is; 113/214(7900X) and 87/161 (1800X). I bet idle power tells the full story, because of extra chip and most likely extra platform features on SKL-X.

 

[Sweepr]

Once again no motherboard model or BIOS details provided, and we already know from Hexus that results vary:

At the time of writing we can only put this down to a lack of software maturity. In the interests of full disclosure, readers should note that the Core i9-7900X initially scored just 4,015 in the VRMark test and the result climbed to 10,191 courtesy of a new motherboard BIOS. There's clearly still work being done to optimise performance.

Intel should just lift the NDA and let the proper reviews (that are ready) be published right now, to avoid any possible bad press from these leakers.

 

[R0H1T]

Yes, it is; 113/214(7900X) and 87/161 (1800X). I bet idle power tells the full story, because of extra chip and most likely extra platform features on SKL-X.

You picked the least efficient 8 core Ryzen, as compared to say the 1700 or 1700x deliberately?

That's why I said TR might be different, but the regular Ryzen is probably the most efficient octa core out there.

 

[tamz_msc]

Yes, it is; 113/214(7900X) and 87/161 (1800X). I bet idle power tells the full story, because of extra chip and most likely extra platform features on SKL-X.

Doesn't look good compared to 6950X. Looks even worse compared to 1700 and 1700X. With Ryzen you get the choice of better efficiency. Perf/watt is an important metric, and total power consumed during runtime is also important in video encoding.

Lets say a 30 min encode @30FPS.

7900X = ((30*60*30)/113)*214)/3600 = 28.4 Wh
1700 = ((30*60*30)/76)*113)/3600 = 22.3 Wh

 

[coercitiv]

If we don't get some more reviews until Monday... this forum will feel like a popcorn bag in the microwave oven for the next 2 days.

 

[SAAA]

Power Consumption

Performance tests looks weird, mixed between case.

But 7900X 10 cores at 4GHz (default all core turbo) consuming around the same as 1800X eight cores at 4GHz is definitely good to see. Both idle and load, quad channel vs dual too.

 

[LaikaSpaceCat]

Looking at the gaming results, which is more sensitive to memory than most other consumer applications, the reduced L3 might be a problem.

Memory latency took a hit, which is the more likely reason for gaming issues than L3 IMO.

It is now between Zen and Skylake in terms of memory latency.

 

[Zucker2k]

You picked the least efficient 8 core Ryzen, as compared to say the 1700 or 1700x deliberately?

That's why I said TR might be different, but the regular Ryzen is probably the most efficient octa core out there.

This is a decacore with high clocks and those numbers, and in the same ballpark as the flagship Ryzen, which is clocked lower. All things considered (platform and firmware), this chip is already looking good, in all areas.

 

[tamz_msc]

Performance tests looks weird, mixed between case.

But 7900X 10 cores at 4GHz (default all core turbo) consuming around the same as 1800X eight cores at 4GHz is definitely good to see. Both idle and load, quad channel vs dual too.

One is at its limit while the other isn't. Not exactly apples-to-apples.

 

[R0H1T]

This is a decacore with high clocks and those numbers, and in the same ballpark as the flagship Ryzen, which is clocked lower. All things considered (platform and firmware), this chip is already looking good, in all areas.

With high clocks & higher temps, the IPC gains (if any) are debatable atm, though over the last decade we haven't seen IPC regression from Intel so I'd wait for more in depth reviews.

 

[Zucker2k]

Doesn't look good compared to 6950X. Looks even worse compared to 1700 and 1700X. With Ryzen you get the choice of better efficiency. Perf/watt is an important metric, and total power consumed during runtime is also important in video encoding.

Lets say a 30 min encode @30FPS.

7900X = ((30*60*30)/113)*214)/3600 = 28.4 Wh
1700 = ((30*60*30)/76)*113)/3600 = 22.3 Wh

SKL-X has octacores too. You avoid the 1800X because it exposes the inefficiency of Ryzen when clocked anywhere near 4Ghz. Meanwhile, the 7900X exceeds that and has more cores to boot. It surely can't be the most efficient chip in the lineup. Let's see how Ryzen R7 fares against 7820X at 4Ghz.

 

[tamz_msc]

SKL-X has octacores too. You avoid the 1800X because it exposes the inefficiency of Ryzen when clocked anywhere near 4Ghz. Meanwhile, the 7900X exceeds that and has more cores to boot. It surely can't be the most efficient chip in the lineup. Let's see how Ryzen R7 fares against 7820X at 4Ghz.

You get the 1800X to ensure 4GHz OC, while you get the 1700 to do 24x7 tasks efficiently. For that you don't need 4GHz.

There has to be a compromise when you chase after the best performance - in this case its efficiency.

 

[SAAA]

One is at its limit while the other isn't. Not exactly apples-to-apples.

That's true, but the opposite can be said if people look at 3GHz lower end skus (say R7 1700) against 4GHz chips to compare performance/efficiency, obviously at lower clocks chips are more efficient. Well I hope for a apple to apple review some day with all cpus at 3-3.5 to chart IPC, performance and power.

Coming back to the current results... cinebench at least shows higher IPC: 2187 vs 1710 and clocks should be 4 vs 3.4 GHz, normalize and it's ~9% more.
Gaming is just too volatile to tell what's happening, but I hope it will follow synthetics later on.

 

[csbin]

That's true, but the opposite can be said if people look at 3GHz lower end skus (say R7 1700) against 4GHz chips to compare performance/efficiency, obviously at lower clocks chips are more efficient. Well I hope for a apple to apple review some day with all cpus at 3-3.5 to chart IPC, performance and power.

Coming back to the current results... cinebench at least shows higher IPC: 2187 vs 1710 and clocks should be 4 vs 3.4 GHz, normalize and it's ~9% more.
Gaming is just too volatile to tell what's happening, but I hope it will follow synthetics later on.

10 Core vs 8 Core

 

[SAAA]

10 Core vs 8 Core

Uh?

I was comparing Broadwell to Skylake here:

Not that if you took the 1800X it would change much: 1614/8x10 = 2017.5 (and I assumed linear scaling to a 10 core Ryzen) then 2017.5/3.6x4 = 2242 (to match 7900X clock).

This nets a 2.5% advantage in total performance but IPC is less for Zen because it has higher SMT scaling (~10%?) in this test, so IPC would be somewhat less, closer to Broadwell/Haswell.