Skylake-X (Core i9) - Lineup, Specifications and Reviews!

Wonder what his temps would be like with conductonaut instead. I dropped 25C on my 7700k after delidding.
 
Excellent comparison of Skylake-X and ThreadRipper, core for core and clock for clock. Unlike biased sites that overclock the interconnect on Zen chips, but leave it on stock for Intel. This review overclocks the interconnect on the i9 as well.

Skylake-X IPC is 32% above Zen. Power consumption is only 17% higher.

8379_23_amd-threadripper-vs-intel-core-i9-clock.png


https://www.tweaktown.com/articles/8379/amd-threadripper-vs-intel-core-i9-cpus-clock/index7.html
 
Excellent comparison of Skylake-X and ThreadRipper, core for core and clock for clock. Unlike biased sites that overclock the interconnect on Zen chips, but leave it on stock for Intel. This review overclocks the interconnect on the i9 as well.

Skylake-X IPC is 32% above Zen. Power consumption is only 17% higher.

8379_23_amd-threadripper-vs-intel-core-i9-clock.png


https://www.tweaktown.com/articles/8379/amd-threadripper-vs-intel-core-i9-cpus-clock/index7.html
What is the price difference between them? Price/performance?
 
I have the 7960X coming in today. If anyone wants to see something specific run, at any clock speed from 4GHz to 5.7GHz, let me know. :D
 
What is the price difference between them? Price/performance?

Since price is a nonlinear function of performance. The company that targets the lower performance almost always has a price/performance advantage. Zero merit for AMD here...
 
Since price is a nonlinear function of performance. The company that targets the lower performance almost always has a price/performance advantage. Zero merit for AMD here...
So basically Intel loses big time on that. Gotcha! ;)
 
So basically Intel loses big time on that. Gotcha! ;)

The same than AMD's 1800X loses in price/performance to AMD's 1700. Not a mystery here, just a consequence of the laws of physics not being linear... ;)
 
Last edited:
The same than AMD's 1800X loses in price/performance to AMD's 1700. Not a mystery here, just a consequence of the laws of physics not being linear... ;)
I guess if people want to pay double the price for not even close to double the speed. Did you already buy yours?
It's going to take a while for Intel to pick it's pants back up and totally beat AMD in almost everything again.

AMD 1800 loses to AMD's 1700, that is all you can say?! Seriously?
 
@ tweaktown:
Mixing higher is better and lower is better graphs on one screenshot = epic fail.

32% higher performance at a cost of 80% more cash and 17% more watts =
... Win??
 
SKL-x has a 24% advantage in productivity despite similiar scores to TR in CB, prime 95, and Handbrake.

Because that extra $800 is worth it for high aida64 and superpi scores.
 
geekbench 4

http://hwbot.org/benchmark/geekbench4_-_multi_core/rankings?start=0#interval=20#start=0#cores=8

23% performance advantage 8 core vs 8 core overclocked, even 5960x's are beating ryzen still.

x265 encoding, this won't even be fair because avx rips through this on sky-x

http://hwbot.org/benchmark/hwbot_x265_benchmark_-_1080p/rankings?cores=8#start=0#interval=20

41.5% performance advantage i took my 7820x score (71.14 fps @ 4.5 ghz due to avx load its reported 4.8 however because that is peak clock) sense its basically the highest non ln2 score vs the highest non ln2 ryzen 7 score (50.28 fps @ 4074 mhz)

there's really nothing that ryzen scores better on be it due to clocks or ipc it ends up loosing almost everywhere and hence why it is not popular on hwbot. (well nasty cold bugs are another reason too but neither here nor there) i think one of the few places it usually does really good on is 3dmark physics scores but overall scores get wrecked every single time due to far worse combined scores. for me sense i bench on hwbot sky-x was an easy platform choice over ryzen even if it costs more with ryzen you can't compete unless your competing against other ryzen cpus and that is why a lot of comps lately have had a ryzen stage.
 
@ tweaktown:
Mixing higher is better and lower is better graphs on one screenshot = epic fail.

32% higher performance at a cost of 80% more cash and 17% more watts =
... Win??


It's a win if top performance is your goal. Not everyone is an "it's good enough" shopper.
 
If you want top performance, you buy a 1950x AND a 7700k system for the same price as a 7960x. That way the price variable is dropped.

TR/KBL wins across the board. No need to settle anywhere.
 
Last edited:
Frito, what exactly are you trying to show us on those grotesqe hwbot links?

I know this is a SKL-x thead, but the last page has been focused on 16 core parts.
 
Frito, what exactly are you trying to show us on those grotesqe hwbot links?

I know this is a SKL-x thead, but the last page has been focused on 16 core parts.
It's called epeen. Ryzen will play games just fine and crunch too. Yes, Intel has the top performance. But price/performance is not as good. There are some Intel's that do have good price/performance.

Like it has been said if you want the fastest possible with no regards to price, then get the Intel.
 
geekbench 4

http://hwbot.org/benchmark/geekbench4_-_multi_core/rankings?start=0#interval=20#start=0#cores=8

23% performance advantage 8 core vs 8 core overclocked, even 5960x's are beating ryzen still.

x265 encoding, this won't even be fair because avx rips through this on sky-x

http://hwbot.org/benchmark/hwbot_x265_benchmark_-_1080p/rankings?cores=8#start=0#interval=20

41.5% performance advantage i took my 7820x score (71.14 fps @ 4.5 ghz due to avx load its reported 4.8 however because that is peak clock) sense its basically the highest non ln2 score vs the highest non ln2 ryzen 7 score (50.28 fps @ 4074 mhz)

there's really nothing that ryzen scores better on be it due to clocks or ipc it ends up loosing almost everywhere and hence why it is not popular on hwbot. (well nasty cold bugs are another reason too but neither here nor there) i think one of the few places it usually does really good on is 3dmark physics scores but overall scores get wrecked every single time due to far worse combined scores. for me sense i bench on hwbot sky-x was an easy platform choice over ryzen even if it costs more with ryzen you can't compete unless your competing against other ryzen cpus and that is why a lot of comps lately have had a ryzen stage.

AVX anything should be tossed as almost nothing other then a benchmark uses it and even in servers it's use is almost non existent. In overclocked test of course Intel will do better as they clock higher nothing anyone disputes. Stock to stock where 95% of the world runs at they are very close. Dont forget to calculate in that cost of a delidder to hit 4.8 as well on that Intel chip. Benchmark junkies are the only ones that care about absolute top score and pay for it, the rest of us are good with close enough and we have extra cash in our wallet for the far more important video card.
 
Frito, what exactly are you trying to show us on those grotesqe hwbot links?

I know this is a SKL-x thead, but the last page has been focused on 16 core parts.

i was gonna compare 16 core to 16 core but sadly nobody is benching threadripper hardly at all. far too small of sample size to see what is typically possible on water.
 
AVX anything should be tossed as almost nothing other then a benchmark uses it and even in servers it's use is almost non existent. In overclocked test of course Intel will do better as they clock higher nothing anyone disputes. Stock to stock where 95% of the world runs at they are very close. Dont forget to calculate in that cost of a delidder to hit 4.8 as well on that Intel chip. Benchmark junkies are the only ones that care about absolute top score and pay for it, the rest of us are good with close enough and we have extra cash in our wallet for the far more important video card.

40 bucks for the delidder, happy to pay that now i can delid any intel cpu currently out there as i have both models now. to be fair though i was quite able to do 4.7 ghz and get almost same score without delidding in fact in x265 i only gained a small amount due to delidding and probably had more to do with being able to clock cache higher now sense it runs so much cooler.
 
@ tweaktown:
Mixing higher is better and lower is better graphs on one screenshot = epic fail.

32% higher performance at a cost of 80% more cash and 17% more watts =
... Win??

32% higher performance for 17% more watts means the efficiency is higher.

32% higher performance for 80% higher price is a consequence of the laws of physics being nonlinear.

Yes, it is a clear win. and more when we add rest of advantages of the X299 platfform: Pcie lanes, RAID, better overclocking,...
 
32% higher performance for 17% more watts means the efficiency is higher.

32% higher performance for 80% higher price is a consequence of the laws of physics being nonlinear.

Yes, it is a clear win. and more when we add rest of advantages of the X299 platfform: Pcie lanes, RAID, better overclocking,...
A clear win for AMD, but Intel is still top. They just need to keep pushing each other so the consumers win.
 
32% higher performance for 17% more watts means the efficiency is higher.

32% higher performance for 80% higher price is a consequence of the laws of physics being nonlinear.

Yes, it is a clear win. and more when we add rest of advantages of the X299 platfform: Pcie lanes, RAID, better overclocking,...

LOL, 32% better performance for 80% higher price due to physics...... my god that is just beyond words.
 
LOL, 32% better performance for 80% higher price due to physics...... my god that is just beyond words.
Clearly you failed Intel math..... :)

Yeah, it is not worth it unless you absolutely have to have the most.
 
aHR0cDovL21lZGlhLmJlc3RvZm1pY3JvLmNvbS85LzUvNzIyMjAxL29yaWdpbmFsL0ltYWdlMy5wbmc=


i mean if we're talking price efficiency.

http://www.tomshardware.com/reviews/amd-ryzen-threadripper-1900x-cpu,5222-9.html

overall pretty good review. still not sure what amd was thinking with the 1900X it makes absolutely no sense its like kaby-x only even more pointless sense at least kaby-x cpus can overclock quite a bit better on average vs standard versions. bottom line though is today bang for you buck is the 8700k if you can get one, if you need top multi core performance sky-x, if you render or just wanna get 12 or 16 cores on a budget then threadripper is for you (but not the pointless 1900x that turns into a quad core ryzen if you wanna game mode it)
 
32% higher performance for 80% higher price is a consequence of the laws of physics being nonlinear.

Yes, it is a clear win. and more when we add rest of advantages of the X299 platfform: Pcie lanes, RAID, better overclocking,...

Intel: even more winning with LESS pcie lanes and "raid keys."

Then again, much of that performance boost comes from the ever-popular aida64 and 720p gaming.
 
If you want top performance, you buy a 1950x AND a 7700k system for the same price as a 7960x. That way the price variable is dropped.

TR/KBL wins across the board. No need to settle anywhere.


Nope, top performance if you aren't worried about cost is 7980xe and 8700K. AMD is not top performance no matter how much you want them to be. Take off the red glasses.
 
Nope, top performance if you aren't worried about cost is 7980xe and 8700K. AMD is not top performance no matter how much you want them to be. Take off the red glasses.
I think he means you can buy a 1950 and a 7770k for the price of just the 7980. That is a pretty good win.
 
I think he means you can buy a 1950 and a 7770k for the price of just the 7980. That is a pretty good win.

We're talking about 2 different things here. I'm talking about performance and he's talking about bang for the buck. They aren't the same. AMD is not about top performance....period. He said top performance.
 
We're talking about 2 different things here. I'm talking about performance and he's talking about bang for the buck. They aren't the same. AMD is not about top performance....period. He said top performance.
I meant 8700k. Top games and you get a dedicated cruncher. Although I bet the 1950 plays games just fine, so not really a huge deal.
Yes, top performance which is crappy price/performance would be the Intel 7980.
 
I meant 8700k. Top games and you get a dedicated cruncher. Although I bet the 1950 plays games just fine, so not really a huge deal.
Yes, top performance which is crappy price/performance would be the Intel 7980.


Like I said before.....not everyone is an "it's good enough" buyer. Why is so hard to understand that? If everyone was they would all own AMD "it's good enough" stuff. Obviously they don't!

Here's the bottom line for you "it's good enough" AMD fans......the best performance in PC gaming/computing award goes to NV/Intel. I know it hurts your feelings, but it's the truth. Just deal with it and move on.


Disclaimer: I don't give a shit who holds the above distinction. I buy top performance and AMD isn't it currently.
 
Like I said before.....not everyone is an "it's good enough" buyer. Why is so hard to understand that? If everyone was they would all own AMD "it's good enough" stuff. Obviously they don't!
Ill see your good enough and say they could buy 2 1950's! How bout that! ;)
Now how much performance would that be......

I do understand what you are saying. It's ok. Enhance your calm!
 
you guys might be shocked by this but *most* of the overclockers i hang out with daily on chats are amd guys, guess what else because they need top performance to compete every single one of them has at least one modern intel platform. try as they have and trust me that they have some even via using phase change on ryzen they cannot compete in a lot of areas because top performance matters when competing and amd doesn't compete there. that is one of the things i was trying to point out earlier on above where i linked some performance gap figures and linked hwbot rankings for some benches and 8x cpu config.
 
you guys might be shocked by this but *most* of the overclockers i hang out with daily on chats are amd guys, guess what else because they need top performance to compete every single one of them has at least one modern intel platform. try as they have and trust me that they have some even via using phase change on ryzen they cannot compete in a lot of areas because top performance matters when competing and amd doesn't compete there. that is one of the things i was trying to point out earlier on above where i linked some performance gap figures and linked hwbot rankings for some benches and 8x cpu config.

There has always been benchmark junkies and they get killed by those willing to put LN2 on their system for suicide benchmark runs. I have a phase change cooler that was built by Jinu some years ago, used it all the time for years but I got tired of dealing with moisture as no matter how hard I tired it eventually got the motherboard. I just use a nice custom water loop these days as cranking the wick to the max got old for me. The funny part even back then I still didn't give a damn about where I ranked, just wanted my computer to run as fast and smooth as possible. But yeah Ryzen is not the best choice for Benchmark competitions, but you should have realized that long before when it was obvious Intel still had the IPC lead, the gap just shrank by quite a bit. AMD or Intel both will run your stuff just fine these days the rest just tends to be overblown.
 
oh yeah for sure one of the guys that recently got a phase has killed a bunch of cpus lately. most recently his 1800x died and he's still not sure how. LN2 runs etc isn't suicide, yeah its risky you make a mistake or do something wrong and you can kill hardware easily. really can do the same thing with normal overclocking as well if your dumb enough its just slightly more difficult but a lot of mobos won't do much of anything to stop someone from running cpus at say 1.6+v even without setting an ln2 jumper.
 
oh yeah for sure one of the guys that recently got a phase has killed a bunch of cpus lately. most recently his 1800x died and he's still not sure how. LN2 runs etc isn't suicide, yeah its risky you make a mistake or do something wrong and you can kill hardware easily. really can do the same thing with normal overclocking as well if your dumb enough its just slightly more difficult but a lot of mobos won't do much of anything to stop someone from running cpus at say 1.6+v even without setting an ln2 jumper.

Well to clarify with suicide I mean you cant run it 24/7 like that. Tho LN2 can kill hardware fast if you screw up. But moisture is a serious problem with sub ambient cooling and I just got tired of dealing with it, if your friends want another phase unit let me know I can sell it, not doing much good in my closet, have to check if it still works first tho.
 
It's called epeen. Ryzen will play games just fine and crunch too.

So up to 30% slower is the new definition of "just fine".

AVX anything should be tossed as almost nothing other then a benchmark uses it and even in servers it's use is almost non existent.

Say that to all the companies that purchased Skylake Xeons to utilize the new AVX512 hardware.

http://www.datacenterknowledge.com/networks/intels-xeon-scalable-designed-ground-data-centers

SKL-x has a 24% advantage in productivity despite similiar scores to TR in CB, prime 95, and Handbrake.

Because that extra $800 is worth it for high aida64 and superpi scores.

SKL-X wins in many other productivity workloads not benched by that review.

About pricing, I always enjoy the double standards. Overall the TR 1950X is only 71% faster than R7 1700 in productivity, but it costs 3x more.

1.71/3 = 0.57

Result? Lots of praise for the 1950X. Youself recommended a 1950X only a pair of posts above.

But if Intel offers 10--20% more performance for extra $800 then Intel is not worth... :whistle:

1.1 / 1.8 = 0.61

1.2 / 1.8 = 0.67

So performance/price ratios are only an excuse, because they are applied selectively to attack Intel hardware.

Same happened with games. Double standards. Skylake-X playing games only 10% worse than the best gaming CPU was considered a kind of disaster, but RyZen being 20--40% behind in same games is "good enough for gaming". :rolleyes:
 
Last edited:
LOL, 32% better performance for 80% higher price due to physics...... my god that is just beyond words.

It is not about words it is about equations. ;)

Clearly you failed Intel math..... :)

In post #1254 I said you that both Intel and AMD are bound by the same laws of physics. And I said you that the nonlinearity of those laws is the reason why "AMD's 1800X loses in price/performance to AMD's 1700."
 
Back
Top