Wednesday, September 9th 2020

AMD Announces a Red October: Zen 3 on October 8, RDNA2 on October 28

AMD CEO Dr Lisa Su on Twitter just announced AMD's next-generation Ryzen processors based on the "Zen 3" microarchitecture, and next-generation Radeon RX graphics cards based on the RDNA2 graphics architecture. AMD is promising a "red" October, with next-generation Ryzen "Zen 3" launch on October 8, and next-generation Radeon RDNA2 launch on October 28. We know for sure that AMD is referring to Ryzen and not EPYC, looking at the Socket AM4 MCM animation being used. The teaser picture for Radeon RDNA2 also hints at a new reference cooling solution with large axial fans.

Update 16:54 UTC: In a separate Tweet, AMD announced the Radeon RX 6000 series graphics cards, based on the "breakthrough" RDNA2 graphics architecture.
Sources: Dr Lisa Su (Twitter), Radeon RX (Twitter)
Add your own comment

141 Comments on AMD Announces a Red October: Zen 3 on October 8, RDNA2 on October 28

#51
Nkd
Legacy-ZA
I love how AMD has been very quiet, it most likely signals something great down the line. I really wanted a juicy tidbit to bite on in the meanwhile though. :laugh:
well now that the date is set. AIBs will likely be getting more info. They have been super tight this time around. I am sure we will start getting leaks later this month or early October. Or even by AMD themselves secretly lol.
Posted on Reply
#52
milewski1015
Franzen4Real
On the GPU side I am a little surprised with the silence. I think nVidia's reveal was for the most part very well received and if AMD were fully confident with their original forecasting of price/performance for RDNA2 vs. 30xx that they would have already brought out some sort of counter info to hold off potential 30xx buyers (being that sales start next week). And if they have a grand slam up their sleeves, they would have definitely made it known now to hold off buyers for another month. Perhaps they have had great yields giving them flexibility in pricing, and therefore are holding off to find out real world performance of the 30xx series relative to their lineup (opposed to marketing numbers) so that they can adjust to either match or undercut nVidia right out of the gate. All speculation of course...
You make some good points. I think you might be onto something about them waiting to see what real world performance looks like and adjusting price accordingly. AMD would be silly to make business decisions solely based on claims made by their competitor. Or, conversely, it could be the fact that they don't have anything that competitive. If they revealed non-competitive products reactively, there's nothing to stop people from buying Nvidia right now. However, waiting to reveal details regarding performance keeps people wondering what RDNA2 will offer and could therefore forego people from buying Nvidia without waiting for RDNA2. Although then you could argue that all they're doing is delaying the inevitable...
Posted on Reply
#53
FinneousPJ
Frick
Do you mean the higher clocks? I'm pretty sure AMD caught up with the IPC with Zen2.
AMD is ahead with Zen2. Intel needs 5+ GHz to beat AMD's 4-point-something.
Posted on Reply
#54
Nkd
FinneousPJ
Yes! But wait is that when they plan to publish the specs or the units?
I think plenty of leaks are about to happen now that official Date is out. AIBs are suppose to get the cards now anyways so I am sure leaks will be plenty up to release. Also they probable planned on this announcement later in the month but did it anyways earlier since people wanted something.
Posted on Reply
#55
R-T-B
dragontamer5788
There's something to be said about Intel's superior and low-latency memory controller
It's not the controller. It's that it's all on one die. AMD could do this too, but it would cost a lot and take a lot of Ryzens advantages away.

And the consensus is that on average, Zen2 beats intel Skylake IPC. Yes there are edge cases, but there always are.
Posted on Reply
#56
M2B
Fluffmeister
Maybe AMD are cranking clock speeds as we speak.
And Voltages of course.
Posted on Reply
#57
Nkd
milewski1015
You make some good points. I think you might be onto something about them waiting to see what real world performance looks like and adjusting price accordingly. AMD would be silly to make business decisions solely based on claims made by their competitor. Or, conversely, it could be the fact that they don't have anything that competitive. If they revealed non-competitive products reactively, there's nothing to stop people from buying Nvidia right now. However, waiting to reveal details regarding performance keeps people wondering what RDNA2 will offer and could therefore forego people from buying Nvidia without waiting for RDNA2. Although then you could argue that all they're doing is delaying the inevitable...
I think they will have fairly competitive product stack. They probably want to see the reviews. Looking like 3080 is going to be around 30% faster on average then 2080ti from the leaked benches. But rumor has it the clocks are fairly set now within few % of up or down. I think they might actually have a card that is competitive rasterization but not as fast in ray tracing. Which is fine with me.
Posted on Reply
#58
M2B
Nkd
They probably want to see the reviews.
LMAO. They will just get the cards and test them.
Posted on Reply
#59
dragontamer5788
R-T-B
It's not the controller. It's that it's all on one die. AMD could do this too, but it would cost a lot and take a lot of Ryzens advantages away.

And the consensus is that on average, Zen2 beats intel Skylake IPC. Yes there are edge cases, but there always are.
One die certainly helps, but even in the Bulldozer era when AMD had a unified chip, AMD's memory controller had higher latency than Intel.

If anything, AMD's memory controller got better with Zen / Zen+ / Zen2. Its split up across multiple chips now with Infinity Fabric in the way, and yet its latency characteristics are passable. Its a true marvel of technology. AMD can't beat simple physics, but they've done really well given the restrictions of chiplets and infinity fabric.

---------------

Intel's Ringbus is just really good. I'm sure Intel has been trying to kill Ringbus for the last decade, but they keep coming back to it. Yeah, it doesn't scale well (in theory), but we're up to 10 cores now and Ringbus is still kicking. And with Gen11 / Intel Xe iGPUs... guess what they're doing? Hanging the iGPU off of the Ringbus. Who can blame em? If it ain't broke, don't fix it.

I mean, I guess its somewhat broke (Ringbus will never scale to as many cores as other architectures), but its doing surprisingly well. I don't think anyone expected to see Ringbus live this long.
Posted on Reply
#60
efikkan
Annoucments…
I'm more curious about the release date.
dragontamer5788
I doubt it. Intel's IPC is very far ahead, and memory-latency is always going to be hampered given AMD's architecture. With Intel's Icelake improving IPC (and finally moving to 10nm), I think Intel is moving faster than AMD can catch up.
There is a good chance Intel will keep the lead, but the gap might actually shrink a bit, because even though Sunny Cove have substantial IPC gains, gaming performance will experience diminishing returns with faster CPUs until the bottleneck is practically removed.
dragontamer5788
I don't think the AMD chip needs to be the best at everything, it should be the best at what it is best at, and just "keep pace" with the Intel chips with regards to single-threaded performance.
Exactly, if only most people could understand this.
Posted on Reply
#61
Makaveli
Frick
Do you mean the higher clocks? I'm pretty sure AMD caught up with the IPC with Zen2.
You are right.

Intel wins now due to clock speed advantage and internal latency (cache design).

They are not "very far ahead" in IPC.
Posted on Reply
#62
sergionography
dragontamer5788
Intel absolutely has faster PDEP, PEXT, DIV, and IDIV instructions. I know, because I've measured them personally. In the case of PDEP / PEXT, 1000% faster or more (single-clock PDEP / PEXT on Intel, but over 15 clocks on AMD).

AMD and Intel both have single-clock add / subtract. AMD had the first dual-AES pipeline (supporting two concurrent AES instructions), so AMD's AES IPC was superior to Intel's AES IPC on Skylake.

What instructions, what conditions, what memory? Its actually very complicated. If we're sitting around adding/subtracting numbers inside of registers all day, both processors tie. MAC instructions, I forget from the top of my head... but you get the picture. Its pretty complicated, because all these CPUs have different speeds for every instruction.
Almost every example you gave is an isolated workload. Remember that you are looking at 2 different architectures with different implementations so it's like apples to oranges. Zen2 for example has bigger cache than skylake, and it's designed that way. If you throw the same size of cache at skylake, you probably won't get more IPC, otherwise Intel would've done it already. I like the performance summary slides here at techpowerup because they average multiple workloads which demonstrates the big picture, so I am referencing one below. You notice that a 3700x at 4.4ghz max, is only about 3% behind a 9900k at 5.0ghz, which has a 10-15% clock increase. When increasing power limits on the 3700x it matches it. Same number of cores, lower clocks, equal performance.
Posted on Reply
#63
Makaveli
ShurikN
AMD late to the GPU party once again...
Thank you for this informative post where would we be without you lol.
Posted on Reply
#64
droopyRO
Raendor
r u serious? Red October is called so because of the revolution of 1917. The movie was named because of the whole USSR theme, and thus relating to communist revolutio.
This. And i really dislike this refference :mad:

I hope to get a Ryzen 4600 if the IPC and/or frequency improvment is accurate. I have zero expectations from RDNA2 and i hope i'm wrong.
Posted on Reply
#65
sergionography
Nkd
I think they will have fairly competitive product stack. They probably want to see the reviews. Looking like 3080 is going to be around 30% faster on average then 2080ti from the leaked benches. But rumor has it the clocks are fairly set now within few % of up or down. I think they might actually have a card that is competitive rasterization but not as fast in ray tracing. Which is fine with me.
I think AMD can win this round or atleast have a stronger case even if they competed in the close to 3080 range. Technically that puts them in a similar placement as the 5700xt compared to Turing, but what is different is:
1. They will only be about a month late compared to a year or more
2. Nvidia's performance came at a great power cost, so AMD is fine not having the halo product if they have the more efficient architecture. Remember Fermi? Gtx480 beat hd5870 in performance but had very high consumption. hd5870 was basically the winner of that generation due to that
Posted on Reply
#66
ShurikN
Makaveli
Thank you for this informative post where would we be without you lol.
Sorry my words hurt your feelings.
Posted on Reply
#68
R-T-B
dragontamer5788
One die certainly helps, but even in the Bulldozer era when AMD had a unified chip, AMD's memory controller had higher latency than Intel.
Phenom on the other hand was really low latency in it's time.

Architecture matters too, and bulldozers sucked.
ShurikN
Sorry my words hurt your feelings.
Less about feelings, more about actually adding to the discussion. That'd be nice.
Posted on Reply
#69
Lindatje
ZEN 3 + RDNA2 easy buy. Throwing my 2070Super in the trash can...
Posted on Reply
#70
ebivan
I'm really hoping the new Radeons will be good. But tbh I dont expect it. AMDs GPUs have not been competitive for along time now and I dont expect them to change now.
They always made a big show and somehow could not delive in the end, first they just got too hot (Vega), then they were just too expansive (VII) or they fucked up driver support (Navi). Its time to bring something that is just good!
Dont get me wrong I am not some kind of fanboy I always buy what I belive gets me the best value for money (I just got a Ryzen 3900x this year) but AMD GPUs did not seem to give me that for a long time..
Posted on Reply
#71
lynx29
Legacy-ZA
I love how AMD has been very quiet, it most likely signals something great down the line. I really wanted a juicy tidbit to bite on in the meanwhile though. :laugh:
If I read the tweet correctly we can expect specs and benchmarks tomorrow too? Seems like they want to counter Nvidia
Posted on Reply
#72
EarthDog
lynx29
If I read the tweet correctly we can expect specs and benchmarks tomorrow too? Seems like they want to counter Nvidia
What, in any tweet, said tomorrow? 9/10? I missed something or.......?
Posted on Reply
#74
EarthDog
ShurikN
im guessing this tweet
Im guessing not? That tweet was yesterday night at 838PM ET. Tomorrow is today? AFAIK, they blew their wad announcing the dates?
Posted on Reply
#75
Makaveli
ShurikN
Sorry my words hurt your feelings.
That is your assumption.

A useless post like that envokes pity not anger! I don't know you personally and vice versa there isn't anything you are going to post that is going to hurt my feelings.

How about adding to the discussion at hand instead of thread crapping?
Posted on Reply
Add your own comment