Friday, June 16th 2017

Core i9-7900X Skylake-X Review Shows Up

An Intel Core i9-7900X has appeared for a full review at the site Hexus.net. Spoiler alert, it clocks to 4.7 GHz on all ten cores with relative ease (only taking 1.25 V, apparently, though it racked up nearly 100°C in Cinebench at that voltage).

The review praised Intel's overclocking headroom and general muscle in a mostly positive review. Still, not all is rosy in Intel land. They found performance per watt to not have improved much if at all, criticized the high price tag, and Hexus.net had the following to say about the overall experience:

"X299 motherboards don't appear to be quite ready, there are question marks surrounding the Skylake-X processors due later this year, and at the lower end of the Core X spectrum, Kaby Lake-X is nothing short of puzzling."

It would seem AMD is not the only major chip-maker who can have motherboards ill prepared at launch time, even the mighty Intel may have teething issues yet.

You can read the full review (which is mostly positive, by the way) in the source link below.

Oh, and a special shoutout to our own @the54thvoid for discovering this article.
Source: hexus.net
Add your own comment

247 Comments on Core i9-7900X Skylake-X Review Shows Up

#26
Unregistered
notbI guess the 14nm is already exploited as far as IPC goes. They wanted high clocks, so they sacrificed some IPC (most likely by limiting cache). In the end this CPU is still faster. Isn't that what we want? Fast CPUs? :)

The most important thing in this CPU is that it matches 7700K in single-thread performance.
Yes, more cores is the future, so this CPU is future-proof in the same way Ryzen 7 or Threadripper are future-proof. But this CPU doesn't have any penalty for the present while Zen does.
It will be near the leaders in games and in vast number of single-thread tasks. It'll be just as good in applications that only use 3 or 4 threads.


The power consumption is significant (way to high for me for sure), but not something that we haven't seen before. It'll get better in time (improved node, optimizations).
Based on how Intel usually improves their architecture, it's very likely that in 1-2 years a successor of this CPU will match Ryzen 7 power consumption, while I wouldn't be so sure about Ryzen 7's successor matching 7900X performance...
I expect 10-15+% performance increase from ryzen 2.0 because of higher clocks and improved ipc. Hopefully overclocked speeds will reach 4.6-4.7ghz. Still, 4-4.1ghz is more than enough for me.
Posted on Edit | Reply
#27
phanbuey
Im happy to see AMD do so well... hoping the 7820 and 7800 do better in games but doesn't seem like it.
Posted on Reply
#28
Basard
I love how the new release of any technology these days is accompanied by "Our fastest chip ever!!" as if they were going to make a newer chip that was slower than the old ones.
Posted on Reply
#29
[XC] Oj101
R-T-BFair enough. I wonder if the 7980XE will be soldered, per chance?
It isn't.
Posted on Reply
#31
R-T-B
[XC] Oj101It isn't.
Figured.
Posted on Reply
#32
Zotz
"it clocks to 4.7 GHz on all ten cores with relative ease"...

I understand what you meant to say, but you're giving a false general impression. It's not "relative ease" if it results in a 100C die temp (with, presumably, a decent cooler). If Intel has cheaped out (yet again) and gimped the thing with a poor heat bridge, they shouldn't get a pass on it.
Posted on Reply
#33
dwade
The powerconsumption on the Bit-Tech review is much lower than the Hexus one. Looks like it's an engineering sample on the Hexus review.
Posted on Reply
#34
efikkan
xkm1948OK some quick search give me this. Yep, a HUGE shrink of L3 Cache.

So Intel is actually going backwards in IPC. Whoever made this decision in Intel needs to be fired 100 times.
Skylake-X features a redesigned cache hierarchy. The old L3 cache had a duplicated L2 cache in L3, but since nearly none of L2 data is ever shared between cores, ~90% of this is waste, meaning ~90% of 256kB per core is wasted.

Intel opted for quadrupling the L2, while reducing the L3 and making it non-inclusive, making the overall cache hierarchy more efficient and improve the hit rate. This makes an increase in IPC.
Posted on Reply
#35
NicklasAPJ
xkm1948OK some quick search give me this. Yep, a HUGE shrink of L3 Cache.

So Intel is actually going backwards in IPC. Whoever made this decision in Intel needs to be fired 100 times.

"How are we gonna compete aganist RyZen"
"I dunno, maybe make our processors even worse?"
"Brilliant idea!"

Does Intel's ass control its brain these days?





Same. Really wanna see how Intel PR and fanboys are gonna spin this story.
They going UP in IPC with giveing the CPU more L2 Cache instead.
Posted on Reply
#36
cadaveca
My name is Dave
NicklasAPJThey going UP in IPC with giveing the CPU more L2 Cache instead.
I don't think the right terminology is in use here. :p

Because of Intel cache changes, AMD will benefit when code adjusts.

It's funny, I seem to find myself almost arguing against "popular" opinion in these threads, no matter if AMD or Intel is the main subject.


:lovetpu:
Posted on Reply
#37
R-T-B
Zotz"it clocks to 4.7 GHz on all ten cores with relative ease"...

I understand what you meant to say, but you're giving a false general impression. It's not "relative ease" if it results in a 100C die temp (with, presumably, a decent cooler). If Intel has cheaped out (yet again) and gimped the thing with a poor heat bridge, they shouldn't get a pass on it.
I simply meant the OC was not technically diffilcult to attain, but thank you for the feedback and (valid) point.
Posted on Reply
#38
cadaveca
My name is Dave
R-T-BI simply meant the OC was not technically diffilcult to attain, but thank you for the feedback and (valid) point.
It shouldn't be, when 4.5 GHz is the max stock Turbo 3.0 bin. Technically, that's only 200 MHz.

I digress, Turbo 3.0 is only on one core, and that core is chosen by the factory, so that OC is a bigger achievement than I might seem to insinuate, but that said, Turbo 2.0 is 4.3 GHz, so it's only 400 MHz.

Also, credit to VSG:

www.intel.com/content/dam/www/public/us/en/documents/datasheets/6th-gen-x-series-datasheet-vol-1.pdf
Posted on Reply
#39
efikkan
Turbo Boost 3.0 is maximum sustained boost for two selected cores, tested at the factory. This information is available for the OS scheduler, requiring a later OS featuring this, which is why you'll see the products feature both Turbo Boost 2.0 and 3.0 speeds, depending on software configuration.

Even though OC is quite possible on these chips, it becomes pretty useless when it's already having such high clocks across all cores and even higher boost. For CPUs costing >= $1000, sacrificing warranty, life expectancy etc. for a few percent is just pointless. Most buyers of CPUs in this range are looking for a workstation to last them 5-6 years, so bumping the voltage of the CPU is out of the question. One of the largest achievements since the days of Sandy Bridge-E is that we no longer have to choose between good single thread and multithreading performance, these large chips are really working well on all workloads.
Posted on Reply
#40
phanbuey
efikkanTurbo Boost 3.0 is maximum sustained boost for two selected cores, tested at the factory. This information is available for the OS scheduler, requiring a later OS featuring this, which is why you'll see the products feature both Turbo Boost 2.0 and 3.0 speeds, depending on software configuration.

Even though OC is quite possible on these chips, it becomes pretty useless when it's already having such high clocks across all cores and even higher boost. For CPUs costing >= $1000, sacrificing warranty, life expectancy etc. for a few percent is just pointless. Most buyers of CPUs in this range are looking for a workstation to last them 5-6 years, so bumping the voltage of the CPU is out of the question. One of the largest achievements since the days of Sandy Bridge-E is that we no longer have to choose between good single thread and multithreading performance, these large chips are really working well on all workloads.
It's not a few % in many cases -- with the 7900x it really is a few % and so i agree with you ... but in some cases:



that's a $700 performance difference. If the 7900X did 5ghz easily and stayed cool you would bet that Digital Storm and Alienware would throw pre-sales boom parties.
Posted on Reply
#41
Prima.Vera
IPC performance/core is hilarious on this one. Epic fail!
Posted on Reply
#42
Aenra
Read the reviews.. some of you guys cherry picked from the results according to your.. side? :)

It does look as if the new cache size plays a role, but only at select applications and only when (obviously) comparing stock frequencies. You run those chips near or as high they can go, you're still on top with the 7900X, anyway you look at it.

(Also.. when a potential reviewer comes and posts about the TIM being a non-issue and how paying fortunes does not entitle us to even the mere basics because "it's already enough".. and he says that while knowing this thing can reach 100C.. Anyway. I'm sure we'll have people reading that review as well, so why not i guess).

My thinking right now is that this is a purposefully obfuscated launch, meant to disguise the simple fact that no one wins.

You had a Broadwell-E and wanted to upgrade? More like a sidegrade.
You had a 4core and wanted to upgrade? More like sidegrade.
You had neither and about time for something new? Better? You check it out, you see that in the end, Intel is still ripping you off.

- You can buy a 6900K or a 6950X (older gen) or you can.. buy something that clocks a bit higher, has same lanes (not more), but a reduced cache, meaning if you don't OC it a lot, the older gen is actually better for you, lol...

- Or you can buy something both new and better in everything (it being the point, right?). Wait until October, buy an 18c monster, disable some of its cores; that way (and only that way) you have same/higher cache than the B-Es had and a higher clock (hence my saying disabling cores, ie paying for nothing. But no other way to have an improvement in both).
That's the only true "better", compared to the previous gen. Anything else is picking what's more important to you and it's a picking that is not even contained within the same multi-core gen.. fail.

And once again, the "true" improvement here costs the typical Intel bucks. 2K-ish for the CPU, 500ish for the mobo.
So in the end, same old, same old. I don't see the price cutting. It's a deception, fooling those (victims) that only care about saying "i have an Intel 8core, not an AMD one". Those that care about the things we care? We're where we've always been, prices-wise.
That's my review ^^
Posted on Reply
#43
eidairaman1
The Exiled Airman
Hugh Mungus4.7ghz is less than claimed before though, but temps are at least not too dangerous. I wpuld rather have a nice, cool, efficiënt ryzen cpu than a student alternative to an induction furnace!
Hugh MungusI expect 10-15+% performance increase from ryzen 2.0 because of higher clocks and improved ipc. Hopefully overclocked speeds will reach 4.6-4.7ghz. Still, 4-4.1ghz is more than enough for me.
Dude start using the multiquote button to insert quotes in your posts, it helps keep the forums on topic and organized.
Posted on Reply
#44
Aenra
And don't stick to that 4.7, i am certain the limit's much higher.
Posted on Reply
#46
Aenra
EarthDogIm confused...
In as far as? :)

edit: aaah, i know; you saw me having tagged you, came here, saw nothing, lol.. O.K., my bad. I quoted an excerpt showing your 6950X doing better than the new i9 equivalent, but then i read the reviews myself, edited post and somehow deleted my original one. I've quit smoking, ain't myself lately ^^
Posted on Reply
#47
Unregistered
Zotz"it clocks to 4.7 GHz on all ten cores with relative ease"...

I understand what you meant to say, but you're giving a false general impression. It's not "relative ease" if it results in a 100C die temp (with, presumably, a decent cooler). If Intel has cheaped out (yet again) and gimped the thing with a poor heat bridge, they shouldn't get a pass on it.
One had a nh-d15s!!!!!
dwadeThe powerconsumption on the Bit-Tech review is much lower than the Hexus one. Looks like it's an engineering sample on the Hexus review.
Entire system and just the chip probably.
#48
johnspack
Here For Good!
To be expected. Intel is running scared. Tiny bit of competition. Ryzen v2 will be even bigger competition. Hurry the hell up guys, I need a $500 20 core cpu soon!
Posted on Reply
#49
cadaveca
My name is Dave
Aenra(Also.. when a potential reviewer comes and posts about the TIM being a non-issue and how paying fortunes does not entitle us to even the mere basics because "it's already enough".. and he says that while knowing this thing can reach 100C.. Anyway. I'm sure we'll have people reading that review as well, so why not i guess).
I don't review CPUs, and I don't care if a CPU "overheats" because I know that they have built-in protections that keep them away from anything dangerous, so really, they actually never overheat. As a reviewer, I know better than to worry about it. Anyone telling you anything different is fear-mongering. CPUs, for me are mere tools I use to do other reviews, and I know my tools well.

Remember so many years ago when everyone was saying OC is dead? That Intel killed it? They weren't lying... but understanding what that meant is a complex subject that it seems only few understand. It's completely baffling to me as a reviewer that enthusiasts don't get it, but, then we get reviews like we did today, and all I can say is "oh well".

Why is 100 C a problem? I fail to understand, because I see zero evidence that shows this to be a real problem. You know there are CPUs that don't even shut down until past 100 C ? If this particular number was a problem, you'd think they'd shut down before that... but they don't.
Posted on Reply
#50
Aenra
One thing at a time :)

You have every right not to care about CPU overheating, throttling, global warming, nukes, cooking, or Islam; every right in the book. Care or care not, that's your call.
The problem arises when your subjective, personal opinions get tangled up with facts, ie what a review should usually be all about.

Now before you'll remind me yet again how you don't tend to review CPUs, i will re-re..mind you that this is a matter of mentality. It encompasses everything, it will as such interfere with what would/could otherwise be an excellent piece of work, helpful and informative. Be it for a CPU review which you don't do or for a mobo review, which you do do. Get my point now? Mentality is mentality. I see you having the wrong one, i'm worried. You get to influence others with what you post, i don't.
(hopefully this puts my original post into a better perspective, though in all honesty, i'd not have expected to need expand on it further)

As to the TIM in i9s specifically?
I wanted to say "if you don't get it, i don't know what else to say", but..
- think of the cost to have them soldered, when they are produced in batches of hundreds of thousands, if not more.. almost non-existing.
- for an almost non-existing additional cost, they make a move that pretty much guarantees insane temps (compared to the opposition's), lower lifespan (all that heat) and/or throttling (lower performance than what you paid for) on an expensive as hell CPU (one out of which they make a fortune already, would have even if soldered).
- somehow, none of the above is an issue and we should all say thank you? And if for some reason you wanna go 101% yuppie on me (take it or leave it, mah market) and insist we should?
i) this time around we have competition, and the competition offers it soldered; for a good reason.
ii) when you dub something High End and 'enthusiast-oriented', you should make sure it's actually in spec with what an enthusiast would expect. You ever bought a Ferrari with a bad radiator? "You have money dude, drive it below 100mph or go buy your own radiator and F off!"
No, not how it goes.

Enthusiast does NOT mean "i'm a victim, sell me anything you damn please and i'll pay extra to make it feasible".
Enthusiast means "i pay extra for better quality/higher performance" (throttling is not an enthusiast's thing, 100Cs are also not an enthusiast's spec, as they prohibit OCing).

As usual, just my opinion, i know in advance most disagree, no worries there :)
Posted on Reply
Add your own comment
Nov 19th, 2024 00:43 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts