• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

Intel Skylake / Kaby Lake

Page 503 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

biostud

Lifer
Feb 27, 2003
15,304
592
126
So maybe intel would have liked to wait with the launch until Q3-4 but with the successful launch of ryzen felt obliged to push X299 platform forward, even with a buggy bios?
 

lolfail9001

Golden Member
Sep 9, 2016
1,056
353
96
Just 20% try again o_O
We are comparing 3.9-4Ghz 1700 against 4.5-4.8 7820X, not stock 1800X and 7900X, thought it was obvious from context as well.
By the way, what load is running?
And costing 500 dollars less too boot while consuming 20-30% less power. I think I know the winner.
And being slower to an extent where just getting job done faster will save me 500 dollars. Not clearcut at all, but we both know where your bias lie. I have already said it, i would not buy SKL-X even if it had 50% rebate, simply too much heat and if i wanted to settle for low clocks, Ryzen would be cheaper until SKL-X gets a ridiculous rebate with it.

Ruining such a high tech piece of hardware with that TIM.....
Pfft, they ruin Xeon Phis with TIM, SKL-X is child's play.

Well, since DDR4-2666 is the fastest the Intel chips can handle, overclocked or not (I am pretty sure) and AMD supports 3000 or better, I would guess the 3000 wen to the Ryzens
Overclocked, Intel handles 3866C11. Needless to say, Ryzen does not come close, but w/e, it is Intel thread, isn't it.

That said, overall, i am not too impressed, feels like a typical pipeline clearance so far for Intel, coupled with corny marketing as usual.
 

AdamK47

Lifer
Oct 9, 1999
13,502
910
126
I really don't like how AnandTech gave Intel a free pass with the "buggy" motherboard BIOS issue when running games. That lack of integrity has greatly diminished this sites standings. Be honest and show us what the performance is like at the time of review.

Such an incredible shame.
 

Qwertilot

Golden Member
Nov 28, 2013
1,584
243
106
I really don't like how AnandTech gave Intel a free pass with the "buggy" motherboard BIOS issue when running games. That lack of integrity has greatly diminished this sites standings. Be honest and show us what the performance is like at the time of review.

Such an incredible shame.
Actively coming out and saying that they can't provide remotely reliable gaming results right now because the bioses are a ****** mess is not giving anyone a free pass in any sense.
 

Bassman2003

Member
Sep 14, 2009
94
14
71
It's looking like Silicon Lottery will be offering delidded Skylake-X CPUs at competitive prices (except for binned chips) with a 1 year warranty. If that comes to pass, it'll will be the best option for enthusiast overclockers, IMHO. For Kabylake-S, the overclocked temps came down 15-25 degrees C!
I agree. The 7900x is a good value if you have the option to get all cores up to 4.5GHz and keep it cool.
 

lolfail9001

Golden Member
Sep 9, 2016
1,056
353
96
Wow. I'd have to watch the video review but that seems like obvious bias. There is no way the 7900X offers better bang for buck than an 1800X. At twice the price, there is no way the 7900X offers better than twice the performance of an 1800X.
If you are running AVX-512 FMA3 loop at 3.8Ghz (as you can), then 7900X offers about quadruple the performance. So... it is theoretically possible?
 

AdamK47

Lifer
Oct 9, 1999
13,502
910
126
Actively coming out and saying that they can't provide remotely reliable gaming results right now because the bioses are a ****** mess is not giving anyone a free pass in any sense.
Yet it's able to run everything else. Run the gaming benchmarks, explain the performance, and revisit it at another time with potential fixes. Intel bumped up the release to counter AMD and in doing so had this issue as a result. Intel should be held accountable with the facts presented.
 
  • Like
Reactions: DarthKyrie

Saylick

Senior member
Sep 10, 2012
931
698
136
If you are running AVX-512 FMA3 loop at 3.8Ghz (as you can), then 7900X offers about quadruple the performance. So... it is theoretically possible?
Hm, fair point. They should've weighted that particular benchmark very low then to reflect how little real world applications actually take advantage of AVX-512 then (the weight should be like 1%).

Side point: I should probably watch the video in it's entirety first before making more claims about the video. :p
 

JoeRambo

Senior member
Jun 13, 2013
978
747
136
What is not logical and completely not OK to me is misguided marketing segmentation and grave errors Intel committed in designing Skylake server/HEDT CPU.

1) Whole 512bit vector effort is very misguided, Intel should have went AVX128 style - use WHOLE set of AVX512 featured instructions on server/HEDT - masks, 32 YMM registers, CD, byte ops. But use them ON 256bits, keeping execution units the same, cache paths the same to save area and power and efficiency overall! This whole effort is full retard mode for Intel and result of what is probably complete loss of guidance inside the company. They already got CPU that does AVX512 real well - that is KL. No one in their sane minds will buy high core Intel CPUs for AVX512 throughput work - it is not power efficient in scale and it hasn't got teraflops of Kl/GPUs either. This whole belief that there are "workstation" or server workloads that can benefit from AVX512 wide vectors is hilariously out of place in the real world where it takes very specific CPU workload: think about being a hard rock of finding 512bit chunks of vector work and hard place of GPUs/KL accelerator that offer way more TF per $ and w. This stupid race to larger vectors on x86 CPUs has to stop, we need actual 512bit vectors in typical server/desktop about as much as we need native 128bit integer registers. Sure some workloads could benefit from those (big integer math stuff, crypto etc), but those are better served by accelerators or dedicated HW like encryption engines.

2) By betting the farm on 512bit execution and wide hw and datapaths to support it ( load bw that is sitting idle if vector units are not in use ) and using new fabric and cache architecture to scale to the moon and feed AVX512, Intel has ruined what were amazing server CPUs. One can only imagine how disastrous memory latency is going to be in 6 mem channels with ECC memory. And Intel made sure it is going to get hit a lot by making L3 smaller and eviction only. Needless to say L3 cache hits were good for power and performance in workloads i care about (think about JVMs with gigabytes of garbage collection per second per server ).

3) This whole local cache L2 size vs L3 is again forced by the need to feed AVX512 monster and scalability to higher core counts. Problem is that not every workload is HPC floating point calculation working on a matrix of numbers. Real world apps tend to have plenty of inter thread communication ( consumer/producer, lock contention, false sharing etc, interrupts, DPC happen) and Intel just made sure they will get penalized properly.

4) On topic of fabric, if latencies are this bad on LCC, what will happen with MCC with 18 cores, even more pain from extra column of cores adding to latency and inter-core traffic jams? At least with Broadwell one could use NUMA on ring, for example run 4 JVM instances on dual socket system each bound to NUMA node and be happy about predictable performance. Same with VMs or containerized stuff like memcached and friends.

5) I don't care about HEDT PCIE lines, as SLI is dead, so GPU + having some extra for two M.2 drives is just fine. Gaming does not care about 8x, compute cares even less. But this whole segmentation leaves bad taste. And this AVX512 segmentation? Even more hilariously out of place, squandering advantages(even if tiny) when none can be spared. Sure I can still develop and run code, but it is spit in the face to segment so obviously - 10 core has it, 6 and 8 cores don't.


Intel still has IPC advantage, but as recent SPEC INT targetted "leaks" have shown us, AMD has great core, with decent power efficiency and 9X% of servers in the world, proliants and poweredges run INT workloads.

Here's a warning word for Intel,
11:54pm up 4285 days, 3:39, 1 user, load average: 1.45, 0.77, 0.64
i ran Opterons in the past and as you can see above one is still running. And I won't hesitate to buy more. Those cents saved on solder, fused memory channels and AVX512 half rate combined with above madness is not OK.

P.S. Disclaimer, i run 7700K 5Ghz on desktop and AMD won't make me give up performance.
 

PaulIntellini

Member
Jun 2, 2015
58
4
71
- AVX512 really seems like a BIG mistake... maybe already AVX256 / FMAC was.
- new cache hierarchy seems ok, but not worth the effort
- mesh is good for the future and for higher core counts, I support making that change.
- IIRC with the 3-memory-channel in Nehalem-EP, some RAS features only worked in downgrade a 2-channel operation mode. Maybe they should have gone to 8 channels all the way or stay at 4.

Maybe the server benchmarks will be better, but this looks like the first bad design mistake since Prescott or mabye Haswell :)
 

TahoeDust

Senior member
Nov 29, 2011
557
404
136
Welp...ordered my 7820x and Auros Gaming 7. This is my first highish end Gigabyte board. My OCD just couldn't deal with the vertical m.2 mount on the Asus boards.
 
  • Like
Reactions: Zucker2k

Topweasel

Diamond Member
Oct 19, 2000
5,339
1,528
136
- AVX512 really seems like a BIG mistake... maybe already AVX256 / FMAC was.
- new cache hierarchy seems ok, but not worth the effort
- mesh is good for the future and for higher core counts, I support making that change.
- IIRC with the 3-memory-channel in Nehalem-EP, some RAS features only worked in downgrade a 2-channel operation mode. Maybe they should have gone to 8 channels all the way or stay at 4.

Maybe the server benchmarks will be better, but this looks like the first bad design mistake since Prescott or mabye Haswell :)
This certainly isn't a Prescott. As a platform this release is close to the 820 release. Needlessly complex and missguided. But this is just a furthering of what we already knew about this platform. These CPU's were made for one thing and one thing only. Servers. Because of that almost all the focus is on Server based performance in a variety of settings ones that even most servers don't see. That way they don't need to have different versions of silicon. I can say with all honesty it probably makes no difference if they disabled AVX512 on the whole lineup instead of disabling a couple units on the cheaper CPU's. But the point of the matter is Intel has never and will never create a true HEDT CPU. The closest to that is the 7700 but their segmentation now means not having more than 4 cores on the performance desktop market. Later it will be 6. Which sadly locks out other people who want performance CPU's. This market and these offerings are all about getting extra margins about taking CPU's designed for one job and putting an i7 and now i9 sticker on them. It didn't matter much in the past because they rarely did anything different between the hierarchy between the desktop and server designs. This changes that. It's still good for people who need ST performance to go with more cores. The platform itself is a mess I would want to stay away from. But these CPU's were never the best at desktop enviroment stuff and its not like this really is that much of departure (cept using the stupid paste).
 

Zucker2k

Golden Member
Feb 15, 2006
1,318
678
136
Welp...ordered my 7820x and Auros Gaming 7. This is my first highish end Gigabyte board. My OCD just couldn't deal with the vertical m.2 mount on the Asus boards.
The 7820x is a beast. Careful, that's one fast system you're about to build.
 

Edrick

Golden Member
Feb 18, 2010
1,914
200
106
Welp...ordered my 7820x and Auros Gaming 7. This is my first highish end Gigabyte board. My OCD just couldn't deal with the vertical m.2 mount on the Asus boards.
I know your pain! I was thinking the same thing today when trying to decide on a MB for my 7820X. But then I realized that I dont want my m.2 drives on the DMI bus anyways, so I will be buying this: https://www.asus.com/us/Motherboard-Accessory/HYPER_M2_X4_MINI_CARD/ so I can use CPU PCIe lanes instead.
 

ASK THE COMMUNITY