• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
  • The Politics forum has been nuked. Please do not bring political discussion to the rest of the site, or you will be removed. Thanks.

RedGamingTech - More PS5 "very unverified" info: Zen 3, SmartShift, Geometry Engine

martino

Member
Apr 25, 2013
6,840
4,834
810
France
RGT is a clout chaser.
He is trying to build his channel, and make some $$ off it, so he pushes fanboy stuff to get those clicks.
Reality is he will get more clicks from a PS5 speculation video than one talking about a new motherboard.
We know from a legit Sony engineer that PS5 doesn't have machine learning.
We know that MS added 4 and 8 bit Int on their own to increase ML abilities, and AMD has not shown the same on their own RDNA 2 cards.
So if he is pushing that PS5 has 4 and 8 bit Int then he is clowning.
We know from the same Sony engineer that the PS5 wasn't full RDNA 2, and that was confirmed by MS saying XSX/S were the only consoles that were full RDNA 2.

Why people refuse to believe what a actual Sony engineer says, but swallow the RGT leaks is beyond me.

Now I am sure Sony made their own tweaks to the GPU and CPU, and these may eventually prove to be better than what MS did, but at this point we have nothing to show for it, so we just wait until it all leaks out, as it will do.
i think you're not understanding brand warriors mindset :
 
Last edited:
  • Like
Reactions: Smoke6

Md Ray

Member
Nov 12, 2016
3,876
12,481
735
so you ignored the rest of what I said there?
a standard Zen2 is more than enough for 60fps across the board
I didn't ignore it. I agree with you.

Anyway, it's an unverified source so it could be BS all along. 🤷🏻‍♂️
 

Boglin

Member
Jul 26, 2013
440
886
625
Rumors and speculation are fun for a lot of people and I'm glad there are a handful of youtubers out there who can scratch that itch. Even if RGT doesn't have real sources of his own it's still a good channel for consolidated rumors.

RGT spent almost two full minutes prior to giving the info that it isn't from one of his usual sources and he doesn't want people to consider the information in this video a leak. Then afterwards, reminded people again that they should take this info with a grain of salt. If people want to take the shit he says after his disclaimer at face value then that's on them if they want to believe in secret, unrealistic features.
For the times he says something is actually true, then I don't even think it matters if he is the first to spread the information as long as he has a good track record for being accurate on the things he verifies or what he considers real leaks.

I'm curious to know from the people who dislike him, is there anyway he would be able to speculate on unconfirmed rumors in a satisfactory manner?
 

Stooky

Member
Oct 18, 2017
410
752
395
I'm sorry for not being more specific as it matters in this context. I do believe he has a source in the AMD pc space as I quoted above. Do you know of any console specific leaks?
You dont know where he gets is info from. the guy has been spot on for the last few months. Damn dude let him live.
 
  • Like
Reactions: Smoke6

danielJackson

Banned
Jun 18, 2020
1,015
2,925
380
Rumors and speculation are fun for a lot of people and I'm glad there are a handful of youtubers out there who can scratch that itch. Even if RGT doesn't have real sources of his own it's still a good channel for consolidated rumors.


I'm curious to know from the people who dislike him, is there anyway he would be able to speculate on unconfirmed rumors in a satisfactory manner?

Some people just lack the imagination to be able to speculate stuff. Everything is black and white to them, and they cant see under the surface.

Also probably because they lack education / knowledge so to them speculations sound like "damage control" or whatever, as they just cant see it as "thinking about interesting matter as it is interesting"

Even this is speculation in a sense and some people cant understand it :messenger_savoring:

Just saying that some people lack the ability to think on abstract level = speculating of the uncertain never satisfies them as it goes over their head
 

Boglin

Member
Jul 26, 2013
440
886
625
FUD post if I ever saw one...

Why FUD? Is it because it said "the GPU sits a max frequency (2.23GHz) 95% of the time, dropping to 2.1 GHz under heavy CPU load, and that happens for less than one frame"?
Seems pretty reasonable to me and I don't think you should be worried about such a minor drop.
 
Mar 7, 2017
2,897
6,105
520
Man, they really prove they have no clue what they are writing.
The 8mb l3 is the same the mobile 4x00 CPUs had. Yes unified is a bit better, that's why the Zen2 CPUs in the Ryzen 4x00 line do not really loose that much performance vs haven a much bigger L3 like Ryzen 3x00 CPUs. But it is basically still the same core and the same on both consoles.

8MB of L3 cache isn't the focus that makes this potentially "big news". 8MB of unified L3 cache is what makes it a potential "big deal" for overall CPU performance.

And you're misinformed. A unified L3 cache drastically reduces the latency penalties of core-to-core communication between the two four-core clusters; effectively turning the 2x four-core cluster (or 4core CCX) of the CPU into an effective single 8 core cluster (CCX). This will have a significant impact in gaming CPU performance specifically, because the biggest killer of CPU performance is latency when CPU cores are sat idle waiting for new work to process. Unlike GPUs, whose massive parallelism makes it easier to hide latency, CPU don't have this luxury and anything that can be done to reduce latency of core-to-core comms, will have a more drastic improvement in gaming performance for the new CPU design.

That's not saying that I necessarily believe RGT's info about PS5 having a unified L3 CPU cache to be true... but it would be one of the most obvious improvements to make for a bespoke console APU design where the scalability benefit of having two separate 4-core CCXs is entirely immaterial.
 

rnlval

Member
Jun 26, 2017
1,296
1,045
460
Sector 001
gpucuriosity.wordpress.com
.

Small correction, it's not '2.11ghz' specifically, it's 2.100's Ghz (hundreds), so it can be 2.190, 2.180 Ghz ect..
95% of 2230 Mhz is ~2118 Mhz and it's used as an example for the clock speed range.

My speculation for PS5 wasn't that tough e.g. 97% to 99% based on RX 5700 XT's behavior between real-world vs paper spec max boost.
 
Last edited:

gspat

Member
Apr 23, 2016
460
236
390
Was actually hoping the feature was SAM.

Mostly because even though windows already has a gimped version, linux/unix/bsd already have a full blown implementation ready to go.
 

longdi

Banned
Jun 7, 2004
8,933
6,710
1,890
RGT is a clout chaser.
He is trying to build his channel, and make some $$ off it, so he pushes fanboy stuff to get those clicks.
Reality is he will get more clicks from a PS5 speculation video than one talking about a new motherboard.
We know from a legit Sony engineer that PS5 doesn't have machine learning.
We know that MS added 4 and 8 bit Int on their own to increase ML abilities, and AMD has not shown the same on their own RDNA 2 cards.
So if he is pushing that PS5 has 4 and 8 bit Int then he is clowning.
We know from the same Sony engineer that the PS5 wasn't full RDNA 2, and that was confirmed by MS saying XSX/S were the only consoles that were full RDNA 2.

Why people refuse to believe what a actual Sony engineer says, but swallow the RGT leaks is beyond me.

Now I am sure Sony made their own tweaks to the GPU and CPU, and these may eventually prove to be better than what MS did, but at this point we have nothing to show for it, so we just wait until it all leaks out, as it will do.

basiscally this.
we can recognise that RGT knows some one from inside Amd or its partner.

But unless Jim ryan personally dm him, there is no logic why he can ramble about Sony PS5 marketing messaging.
The whole L3 cache also dont make sense, you cant just change the cpu isa like a snap. And Sony already mentioned zen2, not zen3, not zen2+.
The boost clocks are the same guesswork from when Cerny said PS5 clocks will drop some here and there.

I hate this gen of YT, collating rumors and guesses, and monetizing it to the fanboys. :messenger_astonished:
 

pratyush

Member
Jun 29, 2020
332
724
275
Another PS5 thread, another war by people who themselves don't know anything more than whatever new source is saying. Read the information and ignore if you don't like it. No need to call the person "fanboy". He has extensively covered Xbox and praised it more than PS5. Problem is people calling him fanboy didn't watch his other videos. It's not RGT who is the fanboy here.
 

rnlval

Member
Jun 26, 2017
1,296
1,045
460
Sector 001
gpucuriosity.wordpress.com
8MB of L3 cache isn't the focus that makes this potentially "big news". 8MB of unified L3 cache is what makes it a potential "big deal" for overall CPU performance.

And you're misinformed. A unified L3 cache drastically reduces the latency penalties of core-to-core communication between the two four-core clusters; effectively turning the 2x four-core cluster (or 4core CCX) of the CPU into an effective single 8 core cluster (CCX). This will have a significant impact in gaming CPU performance specifically, because the biggest killer of CPU performance is latency when CPU cores are sat idle waiting for new work to process. Unlike GPUs, whose massive parallelism makes it easier to hide latency, CPU don't have this luxury and anything that can be done to reduce latency of core-to-core comms, will have a more drastic improvement in gaming performance for the new CPU design.

That's not saying that I necessarily believe RGT's info about PS5 having a unified L3 CPU cache to be true... but it would be one of the most obvious improvements to make for a bespoke console APU design where the scalability benefit of having two separate 4-core CCXs is entirely immaterial.
FYI, the design goals for RDNA to reduce its reliance on heavy forward wavefront stacking to hide latencies. GCN style-forward wavefront stacking is harder to optimize. RDNA has lower pipeline latencies when compared to GCN.
 
Last edited:

Smoke6

Member
Nov 26, 2017
566
615
345
Granted, I didn't take the time to look at every claim he has made, but any claim I have investigated was posted on Twitter or 4chan before he "leaked" it (mostly twitter). Of course, this does leave open the possibility that he may be a primary source on some info, but I have yet to find any situations where that was the case.

I hate to bang on about this guy, I have no grudge against him or anything. Does anyone have any info on "leaks" where indeed he was the primary source?
Nah this is a case of you got caught hating on someone and you didn’t even have the slightest idea who your hating on!

it’s been obvious everytime this topic arises that everyone not a fan of PS will shot on people who have info as if they were part of these so called insiders during the info drought earlier this year!

several posters have stated the guy is reliable with info moreso than anyone else here and y’all just keep sticking to the scripts until that ass just got caught up
 
Mar 7, 2017
2,897
6,105
520
FYI, the design goals for RDNA to reduce its reliance on heavy forward wavefront stacking to hide latencies. GCN style-forward wavefront stacking is harder to optimize. RDNA has lower pipeline latencies when compared to GCN.

Thanks for the insight, but I'm struggling to understand its relevance to my comment.

My comment was really speaking in very general terms about the differences between GPUs and CPUs in relation to their tolerance to latency, but that wasn't even the focus of my post.

The thrust of the discussion was around potential change to the L3 cache on the CPU and the impact that might have to CPU (not GPU) performance.
 
Last edited:

rnlval

Member
Jun 26, 2017
1,296
1,045
460
Sector 001
gpucuriosity.wordpress.com
explains it well.

SiSoftware's Executive Summary:
Zen3 (X, 8-core) is ~25-40% faster than Zen2 (X, 8-core) across all kinds of algorithms. We have to give it 10/10 overall!

Despite no major architectural changes over Zen2 (except larger 8-core single CCX layout and thus unified L3 cache), Zen3 manages to be quite a bit faster across legacy and heavily vectorised SIMD algorithms: it beats the competition even with AVX512 and more cores (e.g. 10-core SKL-X). Even streaming algorithms (memory-bound) improve over 20%. We certainly did not expect performance to be this good.

In effect, it is like getting 50% more cores – 8-core Zen3 performs like a 12-core Zen2 (e.g. 3900X) – and thus even a 10-core 10900K cannot compete. Considering you can just “pop it” into an existing AM4 mainboard (requires a BIOS update to support it) it is a massive upgrade from say, original Zen1/Zen+.
 

MarkMe2525

Member
Jul 16, 2017
1,381
1,752
485
35
Nah this is a case of you got caught hating on someone and you didn’t even have the slightest idea who your hating on!

it’s been obvious everytime this topic arises that everyone not a fan of PS will shot on people who have info as if they were part of these so called insiders during the info drought earlier this year!

several posters have stated the guy is reliable with info moreso than anyone else here and y’all just keep sticking to the scripts until that ass just got caught up
Lol what are you even talking about? Me expressing context to my opinion because I try to be a reasonable human being? I realize certain things are lost when your reading someone's comments, but you are just wrong about my intentions in this post.
 

Smoke6

Member
Nov 26, 2017
566
615
345
Lol what are you even talking about? Me expressing context to my opinion because I try to be a reasonable human being? I realize certain things are lost when your reading someone's comments, but you are just wrong about my intentions in this post.
How were you being reasonable and you made claims before you even researched what and who you even talking about?

once you got an inkling you back tracked that shit!

lesson learned I hope? Let’s just stop criticizing people who are in the know when more than like say 5 or so people on the same thread keeps saying so!

and yet no one has provided any proof claiming he can’t be trusted!
 
  • Like
Reactions: Stooky

MarkMe2525

Member
Jul 16, 2017
1,381
1,752
485
35
How were you being reasonable and you made claims before you even researched what and who you even talking about?

once you got an inkling you back tracked that shit!

lesson learned I hope? Let’s just stop criticizing people who are in the know when more than like say 5 or so people on the same thread keeps saying so!

and yet no one has provided any proof claiming he can’t be trusted!
Guess you got me 🙄
 

Whitecrow

Member
May 7, 2018
2,328
3,772
615
If its "very unverified" why is this being reported?

This why we refer to gaming journalism as "gaming journalism."

:|

Because it's being reported as what it is, without the intention of lying to anybody.
There's nothing wrong on discussing rumors as long as the conversation keeps civil enough, also, this place is a discussion forum (not a journalism website), people come here to talk, not only to inform or get informed.
 

eyesabitdull

Member
May 10, 2020
1,792
3,247
410
Because it's being reported as what it is, without the intention of lying to anybody.
There's nothing wrong on discussing rumors as long as the conversation keeps civil enough, also, this place is a discussion forum (not a journalism website), people come here to talk, not only to inform or get informed.

No no, I get it. And i welcome any sort of discussion, but it does feel a bit too speculative. But you're right, and I retract my statement - carry on :)
 
  • Strength
Reactions: Whitecrow

GymWolf

Member
Jun 11, 2019
19,732
29,866
670
Why in the hell didn't microsoft asked for the same exact stuff on sex??

(If true of course)
 
Last edited:

Whitecrow

Member
May 7, 2018
2,328
3,772
615
No no, I get it. And i welcome any sort of discussion, but it does feel a bit too speculative. But you're right, and I retract my statement - carry on :)
No worries : ) It does indeed feel too speculative.
It would be cool if we got more real info to talk about, instead of 'unverified' leaks.
 
Mar 7, 2017
2,897
6,105
520
It has 76MB of SRAM for the SoC we dont know what level cache is what, but its safe to assume its smaller than a desktop Zen 2 L3
5 + 8 = 76?

MS stated 76MB of SRAM across the entire SoC. There are SRAM arrays everywhere, not just in the GPU and CPU caches. The registers are SRAM. The PHYs for the memory controller contain SRAM. There's SRAM all over the "uncore" and I/O gubbinz that make up the rest of the die.

So alone, "76MB of SRAM in the SoC" is a pretty useless statement. Doesn't really tell us much of anything.
 
  • Praise the Sun
Reactions: Lysandros