Friday, December 27th 2019
AMD Ryzen 4000 Rumored to Offer Around 17% Increased Performance
AMD's upcoming Ryzen 4000 series processors will be based on the company's Zen 3 design, which will feature a deeply revised architecture aiming to offer increased performance (surprising no-one). AMD themselves have already said that Zen 3 will offer performance increases in line with the release of new architectures - and we all remember the around 15% increase achieved with the release of Zen 2 Ryzen 3000 series, which surprised even AMD on its performance capabilities. Several sources around the web are quoting an around 17% increase in performance, taking into account increased operating frequencies of Zen 3 (100 to 200 MHz at least for the enterprise solutions, which could pave the way for even higher increases in consumer-geared products) and increased IPC of its core design. The utilization of EUV in the 7 nm process shouldn't have much to do with the increased frequencies of the CPUs, and will mostly be used to reduce the number of masks that are required for production of AMD's Zen 3 CPUs (which in turn will lead to increased yields).
Sources are claiming an increase of up to 50% in Zen 3's Floating Point Units (FPU) compared to Zen 2, while integer operations should make do with a 10-12% increase. Cores should remain stable across the board - and with that increase in performance, I'd say an upper limit of 16 physical and 32 logic cores in a consumer-geared CPU is more than enough. Increased IPCs and frequencies will definitely make AMD an even better proposition for all markets - gaming in particular, where Intel still has a (slightly virtual) hold in consumer's minds.
Sources:
3D Center, Red Gaming Tech, Reddit
Sources are claiming an increase of up to 50% in Zen 3's Floating Point Units (FPU) compared to Zen 2, while integer operations should make do with a 10-12% increase. Cores should remain stable across the board - and with that increase in performance, I'd say an upper limit of 16 physical and 32 logic cores in a consumer-geared CPU is more than enough. Increased IPCs and frequencies will definitely make AMD an even better proposition for all markets - gaming in particular, where Intel still has a (slightly virtual) hold in consumer's minds.
101 Comments on AMD Ryzen 4000 Rumored to Offer Around 17% Increased Performance
You really have no argument. If we were comparing AMD's Bulldozer, OK I understand your reasoning for buying this Dead End Intel Platform. But today ZEN2 is the King of CPU's regardless what you plan on doing with it.You do realize these Intel processors don't boost to 5GHz when gaming because they are throttled heavily due to heat issues. These processors suck a lot of power and generate a lot of heat.
Maybe the 4000 series will have better luck.
Now factor in that AMD also is the only one with PCIE 4.0 and blazing fast m.2 SSDs it makes buying a z390 and 9900ks a bad deal. Not to mention Z390 is end of line while a X570 motherboard will have at least one more CPU that supports it and maybe two.
I can't see a good reason to buy Intel right now. Even at gaming if you look at higher resolutions you are gpu bottlenecked even with a 2080 ti.
Oh and you are living up to your user name being toxic.
If you shop smart you could have the best.
When is the AMD 4000 series coming out? Around July 2020?
-There are two reasons for Intel's current gaming performance lead, both of which boil down to one point: games are still low-threaded applications. Intel thus wins (slightly) because they have higher max clocks and decent IPC. The other reason is that Intel Core has been the dominant x86 arch for almost a decade (2008-2017 or so), and a lot of game engines are still much better optimized for it. Intel also has higher multi-core turbo speeds, mostly due to ignoring power draw in recent chips.
-The reason the 12- and 16-core chips from AMD perform the best in games for their lineup isn't that they have more cores, but because they have the highest multi-core boost speeds. Games being low-threaded means that AMD's core advantage has little bearing on game performance. It thus stands to reason that a 10-core from Intel won't improve game performance due to core count either. A 10-core Intel chip at the same speed as an 8-core Intel chip on the same arch will perform within margin of error - at least until games start exceeding 4-6 threads.
-The ($750) 3950X trounces the ($600) 9900KS in most CPU-bound applications. Heck, it even beats the ($999) 10980XE in a lot of them. How's that for overpriced? Not all CPUs are made for gaming. And if you need gaming performance, the ($500) 3920X will do you just fine.
-With all of this said, Intel's "lead" boils down to ~11% (unrealistic 720p gaming) to ~1% (GPU-bound 4k gaming), with the realistic sweet-spot of 1440p for a CPU of this price hitting ~3.5% and 1080p for the esports crowd hitting ~6%. Tell me, how does that matter? At all? To anyone?
Why are you ranting about gaming performance? 9900KS is Superior single thread King... End of story
Funny how brainwashed you people are on the subject of Intel CPU Gaming performance... You have every excuse in the book coming out.
I don't even have to upload random pics with quotes trying to discourage and deceive other people into your cult.
9900KS 4K & 1440p gaming reviews right here from the pros of Techpowerup.
4K
www.techpowerup.com/review/intel-core-i9-9900ks/16.html
1440p
www.techpowerup.com/review/intel-core-i9-9900ks/15.html
Sorry that the 9900KS is the King of gaming and hurt your feelings, I'm hoping for AMD to do a better job with the 4000 series.
The only thing I can agree with you FBs is we all benefit from AMD market push, now they woke up the obviously sleeping Giant to build a $7+ Billion dollars super high-end 7nm factory called "Fab42" and there first CPUs coming out on brand new Architecture are Intel Meteor Lake.
Since Intel's Comet Lake is sticking with DDR4, I think AMD's gonna skip DDR5 too.
So I think DDR5 will come with AM5 or AM4+ whatever they call it.I think people were saying in reddit that Jim's intel design will come out in 2023 at the earliest.
Beyond that, first off, the picture I embedded is from that exact review you linked. As were the numbers I quoted. If you actually read the review you might have seen that. To avoid typing it all again, the lead of the 9900KS spans fromIs it the single thread king? Sure. But what does it matter when it's not meaningfully - or, heck, even noticeably - faster than the competition? 6% faster at 1080p means at 200fps the closest AMD competitor would be at 188fps, or at 100fps AMD's option would be at 94. Neither of those are noticeable differences. The differences would drop to near half at 1440p. Is the 9900KS faster? Absolutely. Does it matter? Not really, no. Is it better value for gaming than a 3950X? Absolutely, it's ~$200 cheaper! Is it better value for gaming than a 3920X? No, as it's more expensive. The 3800X or 3700X trounces the 9900KS for gaming value, but they also compete with better-value i7s that make this a bit more complicated. Nonetheless, the advantage of the 9900KS is very small.
Arguing that a 6% 1080p/3.5% 1440p performance advantage makes it "king" ... well, I sincerely hope you can see how silly that is. It's faster, sure. Just not meaningfully, and at the cost of a significant power disadvantage.Agree, but sadly I can't let nonsense like that stand unopposed. Also, using "*girl" in a derogatory way like that is ... quite sexist to say the least. Perhaps you ought to reconsider your phrasing?How would that work? Who would put RAM to the market when there are zero platforms out there capable of using it? These things launch in unison.Yes. I don't think anybody has said differently. AM4 can't support DDR5 anyhow, so a new socket is required for that - which AMD won't be launching until 2021. And DDR5 isn't ready for mass-market adoption yet, so it likely won't be there for LGA1200.
Let me give you a quick lesson, go read a review on the Pentium 4 3.8, and notice how it lost to chips clocked at 2.4ghz. Amd doesn't need clock speed, they need ipc, there is a clock speed ceiling, Intel are hitting it again. They need ipc which they sent doing. If and can boost ipc even 10% then Intel will loose even if they release a 5.5ghz chip. Right now and beats Intel clock for clock by 2-5%, but 5ghz negates that. Now lets make amd 10% faster at the same clock and Intel looses.
Besides, those numbers are for average FPS and not low 1% FPS, which is what matters if you do not want to feel the FPS drop, and if you payed for a an RTX 2080 or 2080TI it would be a real shame to feel those drops.
But can we please stop beating this dead horse? This is not what this thread is about, and just because one troll brought up this stupid discussion doesn't mean we have to keep it going after they got banned from the thread.
2. It may be a dead horse from your point of view, but for me it is not, as the only intensive task I do on my PC at home is game and I do not stream.
3. It is completely relevant to thread as Intel's advantage in the OP was referred as virtual and I was the one who disagreed with it and first brought it up.
Does this 17% IPC increase include the rumored 200-300MHz clock increase, the 7nm+ process enhancement? Or it only reflects the actual Architecture redesign which includes the cache system overhaul.
In my previous post I've asked this similar question.
Depending on what I'm asking, we may be looking at an even higher IPC uplift, over the 17% rumor.
ZEN3 is looking like a very serious contender, much more than the ZEN2 impact. And that ZEN2 was a big impact lol
No because, IPC increase doesn't rule out same / lower clocks. It will however close the distance VS Intel even more and MAY even bypass it, depending on the clocks.
Doubt it because the new process will increase density and therefore higher heat concentration. I actually expect either lower boost clocks or lower clocks in general: something like 100MHz or 200MHz is enough for that 17% IPC uplift to STILL be an performance increase VS Zen 2, in general.
After all, the rumor is about IPC uplift but there hasn't been a peep about performance uplift in general, has it?
Because 7+nm has higher density, it stands to reason it will be more difficult to keep in check, temp wise, and this suggests lower clocks.
Besides, there's this video too (skip to 6:51):
GamersNexus Ryzen 5 3600 review. Look at the AC:O 1440p graph and SoTR. Everything else is very even, the only major outliers are cases where Intel CPUs lag in 1% and .1%.Gaming is the main intensive task for me too, but I nonetheless don't see the point of going for a dead-end platform with zero upgrade path just to get a ~3.5% advantage at the resolution I play at (1440p). That number is from TPU's 9900KS review, btw. That is in no way a noticeable difference, and the tradeoff is getting no upgrade path and a less powerful CPU for everything else - so I'll gladly pass on that.Not really. The news post is about the rumored IPC improvement in Zen3. While IPC is always averaged across different workloads, it would be very hard to reach 17% without also improving gaming performance by a bit. And they only need to improve by a bit to match or beat Intel across the board, even with next-gen Intel chips boosting clocks by another 2-300 MHz. Nonetheless, a post about IPC improvements need not devolve into a rehash of the stale "but Intel is still best at gaming" "yeah, but not by enough for it to matter" "but it DOES matter!" exchange that we've all seen a thousand times.