• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

AMD Ryzen 7000-Series Likely to Launch On or Before the 4th of August

You'd be a fool to buy Intel period. One 5% IPC CPU upgrade and invalid motherboard because Intel wants to force you to buy motherboards with a dozen Intel branded chips whereas you could literally go from some A320 and upgrade all the way to a 5800X3D. Only a fool would willingly give their money to someone who blatantly and known-to-all rips off their own customers.
The reason I've avoided Intel since Sandy Bridge is their security. Sure things like Spectre and Meltdown don't really affect me as a home user in terms of security but the loss of performance due to patches for those does. Buying Intel i would be always afraid of losing performance over time.

Unfortunately barely anyone tests performance with and without patches. Sure AMD does have the same issue but they always seems to be less affected both in terms of number of vulnerabilities and the performance impact of patches for those.

AM4 longevity is just icing on the cake.
 
It's your mony, so it's your choice. All answerss are right. Just don't be bias on that.
This in a nutshell. Trying to be unbiased is the hard part.
 
The reason I've avoided Intel since Sandy Bridge is their security. Sure things like Spectre and Meltdown don't really affect me as a home user in terms of security but the loss of performance due to patches for those does. Buying Intel i would be always afraid of losing performance over time.

Unfortunately barely anyone tests performance with and without patches. Sure AMD does have the same issue but they always seems to be less affected both in terms of number of vulnerabilities and the performance impact of patches for those.

AM4 longevity is just icing on the cake.
Well, I'm still on the beloved sandy-bridge and doing just fine :)

This in a nutshell. Trying to be unbiased is the hard part.
The good news: for most it becomes easier as the years pass by.

You know the saying- with wisdom come age.
Or was it the other way around...
 
So many fools out there...
I guess that make you what? Smart?

I can argue that changing CPU every few years is much more foolish, but I don't.
It's your mony, so it's your choice. All answerss are right. Just don't be bias on that.
I wouldn't say so with AM4. Zen 1 to Zen 2 to Zen 3 were huge upgrades in the same socket. And worth flipping your old cpu for the new one which didn't require a windows reinstall etc.
 
I wouldn't say so with AM4. Zen 1 to Zen 2 to Zen 3 were huge upgrades in the same socket. And worth flipping your old cpu for the new one which didn't require a windows reinstall etc.
Well, my "huge" isn't your "huge".
I'm going from i5 2400 to 13700 or 7700x. That's HUGE.
Also from gtx970 to something like 4060. That's HUGE.

I'm not considering 10, 20, 30 or even 50% performance improvement something worth considering even.
2x and preferably 2.5x performance uplift is what i"m looking for.
 
That really depends on your needs then, anything less than double digits is not worth spending the money on but there are some edge cases where something new (like AVX512 or AV1 hardware based encoding) can help speed up your tasks massively. In that case it could be worth it, basically YMWV!
 
That really depends on your needs then, anything less than double digits is not worth spending the money on but there are some edge cases where something new (like AVX512 or AV1 hardware based encoding) can help speed up your tasks massively. In that case it could be worth it, basically YMWV!
The new Zen4 will have the AVX512 right? Intel removes the AVX512 support from the CPUs. So RLake will not have it? I feel some disturbance and I'm not sure what's Intel's play here. AVX512 exclusives?
I know not a lot of apps use AVX512 and some may say its not gonna bring a lot for AMD but it still helps in some cases though. Never heard of getting rid of a feature or instruction that actually brings performance increase.
 
Last edited:
Not sure about RKL but ADL had it, just disabled by Intel due to the small cores.
It was mostly to save on some power & remove the headache of possible BSOD if a thread jumped from P to E core ~
Part of the issue of AVX-512 support on Alder Lake was that only the P-cores have the feature in the design, and the E-cores do not. One of the downsides of most operating system design is that when a new program starts, there’s no way to accurately determine which core it will be placed on, or if the code will take a path that includes AVX-512. So if, naively, AVX-512 code was run on a processor that did not understand it, like an E-core, it would cause a critical error, which could cause the system to crash. Experts in the area have pointed out that technically the chip could be designed to catch the error and hand off the thread to the right core, but Intel hasn’t done this here as it adds complexity. By disabling AVX-512 in Alder Lake, it means that both the P-cores and the E-cores have a unified common instruction set, and they can both run all software supported on either.
 
Not sure about RKL but ADL had it, just disabled by Intel due to the small cores.
It was mostly to save on some power & remove the headache of possible BSOD if a thread jumped from P to E core ~

Had it or not it is disabled and you can't use it. RL is supposedly not have it at all. It is weird but I guess power saving is the most important topic at Intel and it should be.
I wonder if they will do something about it with the although I doubt it. It will consume more for sure.
 
Last edited:
The new Zen4 will have the AVX512 right? Intel removes the AVX512 support from the CPUs. So RLake will not have it? I feel some disturbance and I'm not sure what's Intel's play here. AVX512 exclusives?
I know not a lot of apps use AVX512 and some may say its not gonna bring a lot for AMD but it still helps in some cases though. Never heard of getting rid of a feature or instruction that actually brings performance increase.
Intel did the mistake of having different ISA support on different cores, which inevitably leads to problems when threads jump between cores. If only they implemented slow AVX-512 support on the E-cores, then this problem would have been avoided. (or skipped the E-cores on desktop of course)
The Zen 4 architecture supports AVX-512, hopefully it will feature this support throughout the lineup and show some impressive performance. It will be especially interesting to see Intel's Clear Linux potentially outperform them with AMD hardware. :)
 
Has it been changed from "recent launch" to "recent announcement"?


Funny that he still says "launch", but the text is "announcement". Something fishy with all this. Unlikely that there is an official launch, I guess.
 
Is it still the same socket or the new socket?
 
Intel did the mistake of having different ISA support on different cores, which inevitably leads to problems when threads jump between cores. If only they implemented slow AVX-512 support on the E-cores, then this problem would have been avoided. (or skipped the E-cores on desktop of course)
The Zen 4 architecture supports AVX-512, hopefully it will feature this support throughout the lineup and show some impressive performance. It will be especially interesting to see Intel's Clear Linux potentially outperform them with AMD hardware. :)
... the whole point is that AVX-512 takes up a lot of die space and cutting it is one of the ways that Intel was able to create their E-cores.
 
... the whole point is that AVX-512 takes up a lot of die space and cutting it is one of the ways that Intel was able to create their E-cores.
If done well, it is possible without much die space. Keep in mind the purpose is compatibility, not top performance. I believe Via did this with their latest design, AVX-512 implemented in a 256-bit SIMD engine, and Intel did similarly with SB with AVX(1) implemented on a 128-bit SIMD engine.
 
If done well, it is possible without much die space. Keep in mind the purpose is compatibility, not top performance. I believe Via did this with their latest design, AVX-512 implemented in a 256 SIMD engine, and Intel did similarly with SB with AVX(1) implemented on a 128-bit SIMD engine.
Neither Via's CHA nor Sandy Bridge are designed with small size and energy efficiency as their top priority.

But go right ahead, keep telling Intel's world-class CPU designers that it's possible within those constraints. I'm sure they'll be fascinated to hear from you.
 
But go right ahead, keep telling Intel's world-class CPU designers that it's possible within those constraints. I'm sure they'll be fascinated to hear from you.
What a wonderful, well founded and convincing argument, which can be used in virtually any discussion, technical or otherwise! It's about as "grown up" as people responding to criticism of athletes with "can you do any better?".

I do believe Intel tried to implement a mechanism where a thread is moved once it runs into a "problematic" instruction, and they shipped this with early firmware enabling AVX-512 for a reason, but evidently this mechanism didn't work flawlessly. If you know how engineering usually works in large companies (whether it's hardware or software), you would know that technical decisions rarely are made by those with the greatest knowledge within the company. The "low-level" engineers are often aware of many bad decisions, so I don't think I have to tell them.
 
But go right ahead, keep telling Intel's world-class CPU designers that it's possible within those constraints. I'm sure they'll be fascinated to hear from you.
AVX512 doesn't take as much space as you think it does, the biggest issue there is still power/heat IMO & with Intel selling "efficiency" cores as a feature it was never gonna work out!
 
Yes, Ryzen 7K/socket AM5 is DDR5 only, but Raptor Lake is not and will work on existing Z690 DDR4 motherboards, also some Z790 motherboards seem like they will launch with DDR4 support as well
I wasnt questioning Raptor Lake tho. The topic was brought up about ddr4 in this thread about am5. I was clarifying/confirming.
 
Except the part about the recent launch then? I mean, this is straight from AMD...
Yes, this specific event is about motherboards, but the text suggests that the CPUs will already have launched ahead of the event.
Well well. We're on the cusp of 4th August here and like i expected there has not been any lauch yet.
Like i said before:
And like i said real launch is always preceded by leaks about final silicon (so far we have seen on ES results), box art, and few days before launch also low quality slides about segmentation (prices, SKU's etc).

I don't see any of these signs yet. So im highly skeptical.
These things need to happen first if if recent news is correct we should start seeing that later this month:
  • Product announcement: August 29, 2022 at 8:00PM ET / August 30, 2022 at 2:00AM CET / 8:00AM TW
  • Press embargo: September 13, 2022 at 9AM ET / 3PM CET / 9PM TW
  • Sales embargo: September 15, 2022 at 9AM ET / 3PM CET / 9PM TW
 
Back
Top