Tuesday, September 22nd 2020
AMD Radeon "Navy Flounder" Features 40CU, 192-bit GDDR6 Memory
AMD uses offbeat codenames such as the "Great Horned Owl," "Sienna Cichlid" and "Navy Flounder" to identify sources of leaks internally. One such upcoming product, codenamed "Navy Flounder," is shaping up to be a possible successor to the RX 5500 XT, the company's 1080p segment-leading product. According to ROCm compute code fished out by stblr on Reddit, this GPU is configured with 40 compute units, a step up from 14 on the RX 5500 XT, and retains a 192-bit wide GDDR6 memory interface.
Assuming the RDNA2 compute unit on next-gen Radeon RX graphics processors has the same number of stream processors per CU, we're looking at 2,560 stream processors for the "Navy Flounder," compared to 80 on "Sienna Cichlid." The 192-bit wide memory interface allows a high degree of segmentation for AMD's product managers for graphics cards under the $250-mark.
Sources:
VideoCardz, stblr (Reddit)
Assuming the RDNA2 compute unit on next-gen Radeon RX graphics processors has the same number of stream processors per CU, we're looking at 2,560 stream processors for the "Navy Flounder," compared to 80 on "Sienna Cichlid." The 192-bit wide memory interface allows a high degree of segmentation for AMD's product managers for graphics cards under the $250-mark.
135 Comments on AMD Radeon "Navy Flounder" Features 40CU, 192-bit GDDR6 Memory
Its a bit of a meme and non-serious, but I think yall get the gist. There's so many conflicting rumors on Navi its hard to take any of these "leaks" seriously.
RX5700 has the same number of cores as the ps5 (2304)
RX5700 has a TDP of 225W
RDNA2 is 50% more efficient than RDNA1
according to AMD
now we can do our speculation/analysis:
If we do the math we can safely assume that 2304 RDNA2 shaders are 50% more efficient than RDNA1 shaders, so that leaves us with 150w at the same performance as rx5700, but in this case; it is clocked around 10-15% higher so that could account for the 175w rather than 150w. Also having cpu cores increases overall package (but not by much since it is clocked low and relies on smart shift technology)
Other things to consider:
Clockspeed for the ps5 is the turbo speed
TSMC 7nm is more mature
AMD is more likely to clock chips higher and sell them as a higher tier product than to clock them low at the sweet spot of performance/watt (this is in relation to this article and my first comment)
Oh right puns and stuff. Uhhh... food, fish, bottom feeding joke or something.... oh right. I hope my joke doesn't fall flat on anyone.
verb (used without object) : to struggle with stumbling or plunging movements (usually followed by about, along, on, through, etc.)
CMO of AMD is on a slippery road
Second thought, maybe it is water cooled. Fury RX.
I don't know about you, but I prefer my leaks not to come from the bleeding, distended rectum of an idiot fanboy.
Also, Ampere IS underwhelming.
1\ Performance scales directly in line with power usage from a 2080 Ti.
2\ It draws buckets of power
3\ It fails when it boosts over 2Ghz
4\ It was a paper launch
5\ And isn't even value for money. You can't get any of these cards for MSRP except founders editions, and there are so little founders editions it makes that essentially irrelevant.
The only redeeming factor of Ampere at this point are they are the fastest cards around. But don't mistake it, this is fermi all over again.
As to the quality of the leaks, I'm quite satisfied with the quality of the leaks and analyses Tom puts out, I've known from him the performance of the Nvidia cards 6 months ago it was right on the money the whole long.
As for the underwhelming aspect, I'm personally not underwhelmed, it's pretty much what I was expecting with the 3000 series. I agree about the hype though, people who believe this is the biggest generational leap are people who have known GPU's only for 4 years. It's a decent generation, who appears huge only because the generation before it was mediocre and expensive.
2. Undisputed - but it puts that power to good use.
3. Anecdotal and uncorroborated. Could easily be PSUs that are not up to scratch.
4. No it wasn't. High demand is not a paper launch. Buy yourself a dictionary.
5. Supply and demand is NVIDIA's fault now?
You're also too optimistic in thinking that 50% perf/W increase is across the board in all cases. The wording was - very deliberately, as this was an investor call - up to 50%. That likely means a worst case vs best case scenario comparison, so something like a 5700 XT compared to the 6000-series equivalent of a 5600 XT. The PS5 GPU with its high clocks does not meet the criteria for being a best case scenario for efficiency. Of course that was stated a while ago, and they might have managed more than 50% best-case-scenario improvements, but that still doesn't mean we're likely to get 50% improvement when clocks are pushed high. All of those fish code names just made me think of Dr. Seuss.
2\ Linear is linear, its the fastest GPU, but its also the most power hungry. Honestly, I expect more from a new architecture + node drop.
3\ Fairly well corroborated between multiple users and now a few in the tech press.
4\ Pull the other one. The tech press knew weeks before hand that there would be limited availability and reported as such. Low and behold.
5\ It's pretty obvious that there is fuck all margin between what Nvidia set the MSRP as and what it sells to AIB's for. Hence only Nvidia's cards are at MSRP and even the cheapest AIB cards are over MSRP.
Its somewhat funny, you blasted in claiming fanboi's everywhere, but man, aren't you the biggest one of all.
3: We still don't know what is causing this or what is happening. Might be a serious issue, might be a tiny firmware fix, might be PEBCAK.
4:
5: Margins are absolutely razor thin, though given just how dense and expensive the FE PCB is, it should be entirely possible for AIB partners to make custom PCB cards at decent quality and sell them at MSRP - they just won't have as good components as the FE. This is a bit shitty from Nvidia's side, but it's better than them setting a precedent for MSPR+$100 as a baseline price like they did previously. That is besides the point, and simply not how you validate your sources. That channel has time and time again presented rumors and garbage data as fact or "leaks", information that has since been shown to have been entirely wrong. That some of what is presented is accurate does in no way make up for this. As the saying goes, even a stopped clock is correct twice a day. MLID is a fundamentally untrustworthy source of information, and anything and everything presented there should be viewed as very dubious. Some of it might turn out to be true, but it's impossible to know what until we have confirmation from somewhere else. Until we can see a significant portion of time when all data presented turns out to be factually accurate, none of what is presented can be trusted to be true. See the GN video above. Stock levels are according to partners the same as or higher than previous launches. Nvidia certainly isn't blameless for the current situation, but this is not a paper launch.
In any case, the part about the memory architecture overhaul is simply an extract from Cerney's PS5 discourse, so it's basically open-source information, not a leak.
Melding your lies with the mainstream's truth in order to make your lies appear truthful is the oldest trick in the book when it comes to manipulating discourse and public opinion (see: Russia and US elections), and unfortunately most people choose news sources based on whether that source agrees with their worldview, rather than how trustworthy said source is. They also have a penchant for doubling down and defending "their" news source when said the credibility of said source is brought into question (instead of holding it accountable), or handwaving the source's inaccuracy away with excuses such as "everyone gets it wrong now and then". Except the dodgy sources get it wrong time and time again.
Make no mistake though, MLID is laughing all the way to the bank with every cent of ad revenue he gets from every chump who watches his reddit clickbait videos. Anyone who wants to reward a liar for his lies, that's your business - but don't expect me to do the same.
But I totally agree his analysis are messy and biased. I'm still waiting for updated PS5 hardware...
The only problem is, your posts reek of bias and ad hominem arguments. So thank you, but no thank you, I value Tom's leaks much more than your opinions, so we'll agree to disagree on this one.
Anyway, back on topic, there are some rumors that AMD has rehauled their memory architecture, but nobody's sure about it, although the people that gave this information also supplied RedGamingTech with real photos of the cards. We'll know more about it in a few weeks.
I am, of course, biased against MLID, as well as RedGamingTech and all those other low quality rumor-mongering clickbait YouTubers. I have so far seen zero reason to trust their content to be anything more than entertainment masquerading as news.
As for the ad honinems: really? Where? Seriously, please show some quotes. As far as I'm aware I haven't commented on you personally whatsoever. (And no, saying your approach to source criticism is poor is not an ad hominem.)