Shintai
Supreme [H]ardness
- Joined
- Jul 1, 2016
- Messages
- 5,678
So #waitforamd again? LOL!
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Skylake-X (Core i9) - Lineup, Specifications and Reviews!
So #waitforamd again? LOL!
DuronBurgerMan
[H]ard|Gawd
- Joined
- Mar 13, 2017
- Messages
- 1,340
If anyone is buying either a Skylake-X or a 16 core Ryzen for primarily gaming... they've lost their minds, or have an awful lot of money to waste for little to no effective gain (most likely both).
Shintai
Supreme [H]ardness
- Joined
- Jul 1, 2016
- Messages
- 5,678
Unlike Ryzen, SKL-X can deliver in gaming.
DuronBurgerMan
[H]ard|Gawd
- Joined
- Mar 13, 2017
- Messages
- 1,340
Deliver what? I'm sure it'll make an appearance at the top of the 3DMark charts. You know, the same place you see absurd 4x Titan Xp SLI rigs and 6950X builds today. But if anyone is buying it for actual gaming experience they'd be better served with a 7700k, or similar. Gaming performance approaches irrelevancy at this level of computing. There is logic to comparing a Ryzen 8 core to a 7700k, because this is a cross-shopped price bracket, and AMD tried to position it as a gaming CPU. This was, IMHO, something of a mistake for AMD. They should have positioned it more as a productivity bargain and, great, it can game okay too if you need that. Then position Ryzen 5 as a budget gaming-friendly CPU (AMD did better with this bracket, IMHO).
At the 10+ core level, gaming is meaningless.
Gideon
2[H]4U
- Joined
- Apr 13, 2006
- Messages
- 3,557
God I love teenage answers to actual insight on how things work.
Shintai
Supreme [H]ardness
- Joined
- Jul 1, 2016
- Messages
- 5,678
Yes that's why we keep having to hear about another fairy tale that never delivers time after time from people like you.
Q2 will even lower AMDs margins, you know with more Ryzen products out, Naples and Vega.
Gideon
2[H]4U
- Joined
- Apr 13, 2006
- Messages
- 3,557
The only one living in a fairy tale is you. Of the three you have mentioned only Ryzen is out. Vega is next then Naples, takes time before they actually show on the books. I mean you actually mention gaming on a 10+ core chip like anyone got it for that reason, sure it will game but will not be the best and a person looking for a 10+ core chip does not care about gaming. Plus there is a 16 core monster coming from AMD that Intel will have to find a answer for as well, it's fun for most of us to see some competition. I guess for fanboys it's a scary time but I look forward to seeing what both sides have and innovation is a good thing.
Shintai
Supreme [H]ardness
- Joined
- Jul 1, 2016
- Messages
- 5,678
So how long time goes it take? I mean its not going to be this year obviously. So what now, 2018?
#waitforamd
#poorvolta
Gideon
2[H]4U
- Joined
- Apr 13, 2006
- Messages
- 3,557
6 months from a product launch to see it on the books is pretty typical.
depends on their release, first full quarter results are always best to look at, but if launched in the beginning of quarter, a product such as Vega, it should be the highest selling quarter, demand tends to be higher when people are waiting for a product, of course if it can be supplied in quantity too.
Are we really thinking AMD is going to do anything with Naples, with its cross talk CXX problems, Naples might have more issues than current Ryzen variants, more CXX modules. Also server side of things, stability is a problem, AMD must ensure system stability before anyone picks them up. There is no place for consumer beta testing. That takes time to get the consumer confidence. So two hurdles, NO CXX latency problems, then no system stability issues even before talking about comparative performance to Intel counter parts.
This is why AMD stated 2017 isn't going to change much for them, they know the hurdles they need to get across before their products are adopted.
This is exactly what I was stating, over 2 Q's ago, its not an easy road back to competitiveness for AMD, doesn't matter the market. They have to prove themselves all over again. And it doesn't help with slide shows and poor launches. They need to dot every I and cross every T if they want their consumers to be willing to take a risk with them. Can't have graphics cards that go over PCI-e spec, can't have 1 q of waiting for motherboard bios to fix memory clock issues. Can't have the CCX problems showing weaknesses in certain applications. All of what they do reflect on what they are capable of.
Are we really thinking AMD is going to do anything with Naples, with its cross talk CXX problems, Naples might have more issues than current Ryzen variants, more CXX modules. Also server side of things, stability is a problem, AMD must ensure system stability before anyone picks them up. There is no place for consumer beta testing. That takes time to get the consumer confidence. So two hurdles, NO CXX latency problems, then no system stability issues even before talking about comparative performance to Intel counter parts.
This is why AMD stated 2017 isn't going to change much for them, they know the hurdles they need to get across before their products are adopted.
This is exactly what I was stating, over 2 Q's ago, its not an easy road back to competitiveness for AMD, doesn't matter the market. They have to prove themselves all over again. And it doesn't help with slide shows and poor launches. They need to dot every I and cross every T if they want their consumers to be willing to take a risk with them. Can't have graphics cards that go over PCI-e spec, can't have 1 q of waiting for motherboard bios to fix memory clock issues. Can't have the CCX problems showing weaknesses in certain applications. All of what they do reflect on what they are capable of.
Last edited:
Shintai
Supreme [H]ardness
- Joined
- Jul 1, 2016
- Messages
- 5,678
No its not.
Archaea
[H]F Junkie
- Joined
- Oct 19, 2004
- Messages
- 11,826
Hey I am looking forward to gaming on my future Skylake-X thanks
And buying an Optane SSD when they launch the consumer version late this year
Probably will be the 8C/16T as it will only have light duties.
Cheers
Intel is making sure any bracket over 4 cores is also theirs, they aren't giving AMD any room to say we have more cores so its a better buy. If people want more cores and the same performance in apps that use less cores, while getting more performance with apps that use more cores, Intel is still the way to go. Of course consumers still have to pay for that extra lol and its probably (if not definitely) not going to be in the same cost brackets of Ryzen.
There is no compromise between what performance people want with the skylake x chips.
DuronBurgerMan
[H]ard|Gawd
- Joined
- Mar 13, 2017
- Messages
- 1,340
Hey I am looking forward to gaming on my future Skylake-X thanks
And buying an Optane SSD when they launch the consumer version late this year
Probably will be the 8C/16T as it will only have light duties.
Cheers
Well then, sir, you're probably in the second bracket. After all, Ferraris have a place in the market too. Some folks have the scratch for 'em.
DuronBurgerMan
[H]ard|Gawd
- Joined
- Mar 13, 2017
- Messages
- 1,340
It's a performance jump, which is nice. But we have more or less the same situation with Broadwell-E already. Ryzen is competitive with Broadwell in most (but not all - there are numerous cases Intel still clearly wins) multi-threaded workloads, but sacrifices relative gaming performance to get there. In exchange, you can build a Ryzen 7 box at a cheap, mainstream price point. Skylake-X vs. 16 core Zens is shaping up to be something rather similar. I expect Zen-based HEDT parts to be significantly cheaper, but give ground to Skylake-X in gaming, and in edge cases. As always, buy Intel if you want the best and can afford it.
I ain't Nostradamus, but I've one prediction: AMD will never win the absolute performance crown again. Not in CPUs, not in GPUs. What I expect out of them is competitive mid-range products with good price/performance, and high-end options that are competitive in specific workloads at significantly lower prices.
If anyone thinks a Zen chip is going to crush Skylake-X outside of very specific workloads (and maybe not even then), they are dreaming the fanboy dreams.
Just give you an example of a none gaming situation with a Ryzen build in the mid range. My mom wanted to updated one of her computers on the cheap at her office that was outdated and using older Windows (with the recent scare of the crypto virus). So I told her its a bad time to get computers, no real sales no real deals, the best she could get for the price she was looking at is Ryzen 5 , she said now, ya know why, platform, not stable, can't take the chance. Ended up finding a lenovo skylake 6700 non k system for 600 bucks SFF on newegg. Ended up getting a better deal for an Intel system than what a Ryzen 5 system could offer.
Its not just performance. Its everything. AMD went down hill for 10 years, and now to get consumer confidence back, its not a price thing, price matters little in business, its only going to come down to the system in its entirety.
For gaming, I see no need for anyone to recommend Ryzen systems unless the consumer is on a seriously tight budget.
Its not just performance. Its everything. AMD went down hill for 10 years, and now to get consumer confidence back, its not a price thing, price matters little in business, its only going to come down to the system in its entirety.
For gaming, I see no need for anyone to recommend Ryzen systems unless the consumer is on a seriously tight budget.
Gideon
2[H]4U
- Joined
- Apr 13, 2006
- Messages
- 3,557
You would never recommend a AMD system, lets be real. You try to act otherwise but your bias is obvious. Also the crypto virus has nothing to do with hardware and everything to do with not clicking on stupid stuff and having the windows update for it as well. If you were a good son that is what you should have told her. Thanks for the laugh tho.
Or use to hobbies that are expensive; try audiophile hobby with high end equipment
More seriously it is down to engineering principles rather than anything else with regards to Skylake-X. along with some workloads and the very last bit of performance possible that needs said platforms.
And that the other bonus (or cherry on top) is getting the best out of games.
I think of the HEDT Skylake-X range the 8C and 10C will be pretty popular.
Last edited:
Software is a problem, the only systems that won't be affected it was Windows 10 or 7 since those were the ones getting updates from MS on a regular basis, that system was still using windows xp.
yeah I wouldn't recommend AMD the only reason I stated it was because of the 500 - 600 price tag she was looking for. Guess what Ryzen systems are cheapo systems, and you get what you pay for a system that is untested is NOT good in business. It wasn't even me saying that to her, it was the medical software company that provides her EMR that stated it to her.
You guys just don't understand how businesses operate, none of these things make sense to you, but there is much more to logistics then just having hardware that is available.
Just to give an example if she gets a ryzen system and something goes wrong, lets say the EMR software crashes on that system. How much money is my mom out? First thing she has to pay the person using the computer more time because that person now has to double up on another person's computer, so that's 20 bucks an hour extra for how ever time that is. Then she has to pay for the EMR company to send a guy over, that is 50 bucks an hour and a service call which is 500 bucks. Now she is a doctor and goes to ER at times, so if there is something she needs to be pulled up in a chart and she can't get to that information in a timely fashion? What happens.
just one hour of her office's time is more than 2k an hour. Now her time is more than 1k an hour. So spending time to find a cheap system than may or may not work well for her needs is a waste of time.
These are simple things that can be avoided. just spending a couple hundred more initially, would save her possibly 1000 bucks or more and without the headaches.
DuronBurgerMan
[H]ard|Gawd
- Joined
- Mar 13, 2017
- Messages
- 1,340
Or use to hobbies that are expensive; try audiophile hobby with high end equipment
More seriously it is down to engineering principles rather than anything else with regards to Skylake-X. along with some workloads and the very last bit of performance possible that needs said platforms.
And that the other bonus (or cherry on top) is getting the best out of games.
I think of the HEDT Skylake-X range the 8C and 10C will be pretty popular.
For audiophile stuff, it makes a lot of sense. I am a producer and a DJ in my other life. Piles of cores are excellent for this. Stacking VSTs eats threads like nobody's business, let me tell you.
It's why I didn't buy a quad this time around. Broadwell-E and Skylake-X are out of my price range -- though I regret that, since I probably would have bought a 6900K if I had the budget for it. Unfortunately, I'm half Armenian, and thus I am probably genetically predisposed to be cheap as f*ck. But as a budget-friendly option for this use, Ryzen was good. More than twice as fast as my old 2600k. Now, if Coffee Lake mainstream 6 core chips were out at the time, that may have been a good option for me as well.
IMHO Coffee Lake 6 core is where it's at for most of us. The things will sell like hotcakes if priced right, I think.
DuronBurgerMan
[H]ard|Gawd
- Joined
- Mar 13, 2017
- Messages
- 1,340
TBH, I have a hard time recommending AMD systems too. And I'm very brand agnostic. Let's be fair to AMD for a moment. They did very good with Ryzen. Architecturally, they leaped forward from circa 2008 to 2014 with ONE release. That's no small feat. And then, top top it off, they managed to spin said 2014-ish architecture into something viable and interesting in 2017 by giving you more cores for the dollar. Again, impressive.
Someone else here said that the problem isn't AMD's engineers. They are able to work some fantastic miracles with comparatively tiny R&D budgets. The problem with AMD since ~2007 or so has been leadership and marketing. Once, AMD actually stood toe-to-toe with Intel. They had good leadership, talented engineers, and marketing folks who knew how to sell them. AMD approached Intel's market share in 2006. 59% Intel/41% AMD, IIRC. But after Athlon X2... their management got stupid. They *intentionally* threw away their presence as a technology leader, and went back to building budget sh*t filed with bugs and antiquated technology. And every marketing hype train departed the station to the same destination: the garbage dump. All the effort spent since K6-2 to become competitive and build a better business... they just said f*ck that.
And me, as a customer... I remember that. A lot of us do. AMD tarnished their own name. And so even if AMD devastated Intel (which it didn't), I'd be somewhat wary of trusting them.
All that being said, I still bought a Ryzen chip. And it's pretty good, if maybe not all it was cracked up to be. Maybe AMD learned their lesson, and is, like during the K6 and early K7 days, catching back up and intends to build serious products again. I hope so man -- I really do -- but you'll have to excuse a lot of us if we're still skeptical, and aren't sure if we're ready to trust them again.
Yeah Coffee Lake 6C as you say is going to have a lot of interest and rightly so, just for others Skylake-X is looking pretty good at 8C and 10C if also interested in both gaming and other workloads and tech solutions.
Cheers
Last edited:
Not true. For instance the nonlinear functional relationship between IPC and number of transistors is a physical law and independent of the existence or not of competition. That is the reason why increasing IPC by 50% on top of Excavator didn't cost 50% more transistors to AMD engineers but many many more.
CPU are different designs than GPUs. Scaling and complexity is not the same for latency than for throughput.
Not even AMD believes that, as their guidelines for the rest of the year prove... not to mention those i9/R9 products are addressing a niche market of a niche market: aka infinitesimal sales.
Napoleon
[H]ard|Gawd
- Joined
- Jan 27, 2003
- Messages
- 1,073
let's hope for the 'trickle down' effect to other CPU segments
(ones that actually sell volumes)Agree about RyZen, but SKL-X is different. The 8-core SKL-X has IPC at Kabylake levels for gaming, moar cores/threads, quad-channel, and 4.5GHz turbo.
Take the 6900K in the next graph
Add 5--10% extra IPC. Add 15% higher clocks and the 8-core SKL-X will score about 190 points, beating the quad-core KBL. Another thing will be if the extra performance deserves the pricing or if gamers will be well-served by KBL and CFL.
Glaring fakes in the threadRipper lineup exposed in the corresponding thread in the AMD section of forums. Let us discuss the real SKL- lineup in this thread.
I hope some day people as you get that reporting facts, telling the truth or correcting people when is wrong is not a synonym for "bias", "trolling" and the rest of the usual arsenal of ad hominems.
Teenyman45
2[H]4U
- Joined
- Nov 29, 2010
- Messages
- 3,244
I wasn't talking about the efficiency and efficacy of the underlying silicon and transistors, but how much the chips be it cpu, gpu, memory, or whatever ends up costing the consumer. When there is meaningful competition, the percentage of performance increase between design A and design B has a corresponding, if not linear increase in dollar/euro/pound price. When there is no competition or when the resulting product is so cutting edge it really can't be easily compared to lesser chips from the same company, then the price to performance relationship breaks down.
From the late 2000's until AMD's release of Ryzen Intel has effectively monopolized the high performance silicon bracket and has charged accordingly. Look at the premium until charges for an unlocked multiplier on mainstream cpus or the ridiculous price of the 10 core 6950x. The longer this has gone on, the worse it has gotten. Nvidia does the same thing with its gpus (See the last few years of titans and the recent 10XX FE series price gouging), and AMD tried without success do it on the 7970 launch which was one of the few times they had a brief shot at the GPU performance crown.
Gideon
2[H]4U
- Joined
- Apr 13, 2006
- Messages
- 3,557
You dont report facts at all, you speculate and tell everyone else they are wrong then change the subject when your wrong. Now since it's a intel a 12 core chip is great for gaming even tho it wont clock up, your bias is pretty obvious when you ran around telling everyone needs just a 7700k when Ryzen was released as ipc and clock speed is all games need. For people that actually want info instead of opinion here is a lineup from both companies on the upcoming chips.
DuronBurgerMan
[H]ard|Gawd
- Joined
- Mar 13, 2017
- Messages
- 1,340
That was my point to begin with. The value proposition doesn't work for a pure gamer. And 7700k already has 4.5GHz turbo. Skylake-X will likely score marginally higher in some cases (we already see that on occasion with Broadwell, and as you say, SKL has better clocks and IPC). But difference will be negligible in most cases. So unless you're willing to pay a *huge* premium for minimal (if any) gain, there's no reason for a gamer to go Skylake-X over 7700k or future CFL 6 cores. Though I'm sure will see the competitive overclockers on the 3dmark leaderboards hit up Skylake-X for maximum benchmark scores.
Now, if you're talking mixed use with multi-threaded applications, streaming while gaming, or pure multi-threaded workloads, the value proposition changes. But I suspect even most of those consumer-level users would either be fine with a Ryzen purchase, or a CFL 6 core. Skylake-X will be the best here, though, for sure. So if someone has the money, it would be worth looking into for those use cases.
You gave a good example for a small business, that wants new but stable. So AMD has a long road ahead on that side. However, the consumer market works differently: the concept in that market is "Is it GOOD ENOUGH". That's where the volume is. So Ryzen should capture volume in that market. Now they have to continue to keep up with Intel. It will be a big distraction for Intel right when they need to be focusing on other stuff (AR/VR/ADAS/etc).
For most people 2 core Intel's are enough. The only reason my mom wanted to stick with 4 core systems was because they do come in handy at times even for regular office use. I recommended i-5 to her over 5 years ago. A bit more money for computers that she doesn't need to upgrade for 7 to 10 years. My mom doesn't like to upgrade her office systems frequently, cuase the costs associated with it, they need to call up their EMR reps to come and put them in which will cost 2 days of work, not cheap at all.
Now for a general consumer, the only reason to get more cores is if they are usable for their tasks, typical upgrading on a home pc is 3 to 4 years? What is the use of getting more cores now, when they aren't usable right now, but in 3 to 4 years a person is going to upgrading they can get it right now. This is why Intel didn't move away from 4 core for mainstream systems, cause the need isn't there, Coffee Lake was slated years ago to go 6 core. AMD Ryzen hasn't changed Intel's view nor software development views either.
You are correct that the extra cores might only provide marginal benefit. But that's where marketing goes in: similar performance, but cheaper, more cores, more more more for less less less. Watch the crowds storm Walmart on Black Friday in the US, and you will get an idea what I mean.
Master_Pain
Supreme [H]ardness
- Joined
- Apr 13, 2007
- Messages
- 4,947
You know that there are games in fact, that do use more than 4 cores, right? Cities Skylines chewed my 4930k up and spit it out. I upgraded to an 1800x for just this reason. And I will probably upgrade again to a 10+ core once the new i9s and R9s hit and I see how they do and make sure they keep the clock speed up.
People like you, insisting that we don't need more than 4 cores for gaming, is what has allowed Intel to shove quads down our throats for so long.
Indeed. Most game producers aren't going to optimize for multi-core systems without knowing that their user base will take advantage of such optimization. Intel, has used this as an excuse to slow down core count increases on consumer chips. They then lock the high core count server chips to prevent anyone from using the very ample headroom in them (so they can incrementally and glacially increase speeds in the future, charging more and more each time). AMD has kicked things in gear again, giving game developers reason to optimize for multi-core systems by lowering the cost of them so that more and more people can get into a system with more than 4-cores.
Intel must be every different shade of pissed off, as AMD has basically kicked Intel's well-planned business model (that milks us to the maximum extent possible and denies us as many features on our CPUs as possible) squarely in the balls, leaving Intel no choice but to increase core counts of their HEDT chips years ahead of when they'd have otherwise done it.
I look at the rumored PCI-E lane count for Ryzen 9 and smile....44 lanes across the board. Intel? Only the top two SKUs have a full lane count, with the latter two Skylake-X SKUs being neutered to 28 lanes. Why? Because that's how Intel operates...nickel and diming you to death on features that are free for them to implement just because they can. Poor Intel, it must really suck for them having to compete again and earn our business, rather than just getting it by default for whatever Xeon scraps they felt like offering us.
DuronBurgerMan
[H]ard|Gawd
- Joined
- Mar 13, 2017
- Messages
- 1,340
Yes, yes, it is all my fault. I used my secret contacts with Intel to convince them to sell only quads at reasonable prices. Even though I built a Ryzen rig. Even though my name contains an AMD processor. Oh yes, it was all part of the plan. I'm secretly in league with Shintai (even though I spend half my time on this forum arguing with him).
In all seriousness, it's not a matter of insisting we don't need them, it's a case of few games being well-optimized for them. Chicken and the egg scenario. It took awhile for them to optimize for duals. Then for quads. Years, my friend... years. We're now at that point where we'll start to see more than 8 threads get utilized (and we've seen it in a select few games). And lo and behold, Intel appears with a 6 core mainstream part right about the same time. If someone insisted on building a gaming rig today, I'd say 7700k. But tomorrow? CFL 6 core parts, if they clock anywhere close to the 7700k and are reasonably priced, will be a better buy. Ryzen, of course, will continue to be the budget multi-core solution.
The point was not to sh*t on Ryzen, but to point out that Skylake-X at its high price point isn't offering you much more for gaming than a Kaby, CFL 6c, or a Ryzen will. I.e., not worth the money from a gaming build perspective alone. Much better value proposition if you're talking mixed-use workstation, though.
I am not talking about efficiency and efficacy, but about performance and cost. As stated above, the laws of physics give a nonlinear relationship between performance and cost, and evidently those laws apply both when there is competition and when there is not.