Radeon VII reveal burries the hopes for trully High-end Next-gen consoles.

lol...you are dumb. Actually, I don't know if you are dumb, trolling, or just do not understand the English language that well. lol This is comical. Please keep going.

k

@blackhairedhero: Im not going to sit and get in an insult war with you. You have shown yourself to be a fool that's not worth further attention.

@Diddies: " To a word that's not real"

That is the proper structure.

For an english expert you struggle.

@fedor: Cool story...

Learn how to type before you criticize others and don't be butthurt if the PS5 is over 10 tflops.

@blackhairedhero: What are you talking about? You can't spell or use punctuation, you came in insulting people, and you used unsubstantiated rumor to push your ignorant narrative. Get lost, troglodyte.

This is funny. You keep misplacing your quotes every time. Come on...learn the English language.

@fedor: I misspelled one word through a mobile device. Meanwhile you butchered an entire sentence and capitalized "Who" in the middle of your sentence. Honestly I don't give a shit about grammar in a forum as long as I can get an idea of what your saying but you wanted to make this the theme, so be it.

You are a clown and unlike you at least I used credable sources. You used nothing other then your own bullshit opinion.

@blackhairedhero: Why are you mad at me? Its not my fault you're a moron.

Link your sources too, troglodyte.

@Diddies: You sound like a Ghetto Hamster.

Misplacing my quotes? You can't quote from mobile.

@blackhairedhero: Use the desktop option in your browser, moron. He's also talking about you misquoting sentences, which you've done several times now.

@fedor: Mad at you? I'm just laughing at you. You came in rabid because I insulted your Tflop prediction and then you had this complete meltdown. It's honestly hillarious to me.

https://www.google.com/amp/s/www.eurogamer.net/amp/digitalfoundry-2018-in-theory-can-a-potential-ps5-deliver-a-generational-leap

Now when you are done having your meltdown come up with something more substantial to back up your claim.

@fedor: I didn't misquote your sentence I just quoted the part that was broken.

@blackhairedhero: You need to read the exchange again. Also, thanks for the link. Once again its just rumor and theory. I'd think you were doing some sort of divine satire if I didn't already know just how uneducated you are.

@blackhairedhero: So you misquoted then. Also, your critique was incorrect. Hilarious.

I mean sure my numbers are off by 5-10% at most, i did it in my head. Go ahead and post accurate numbers

Even with a 2020 launch experts like Digital Foundry expect 12 Teraflop minimum

But what do they know, they just analyze tech and performance for a living. YOU'RE Clearly the expert

Not sure what you mean by Gio

@neutrinoworks: Digital foundry link states clearly its all rumor and speculation. Stop trying to pass it off as confirmed, gio.

@fedor: I did read it and I copy and pasted the part of your sentence that was broken, then I fixed it for you.

I have a Masters degree but it was in international business as opposed to English. But it's ok I can deal with a rabid forum dweller hurling insults at me.

Regardless my link wasn't based off rumor. It was centered around what AMD has accomplished and what's possible according to tech experts. You on the other hand have yet to provide anything other then opinion.

@blackhairedhero: Your link is rumor and speculation. You bore me.

@fedor: Do you have any link's? Because it seems like the only thing you have is wishful thinking.

@blackhairedhero: Yea, I do.

https://www.hookedonphonics.com

You need that desperately.

@fedor: Lol your meltdown is hillarious. Nothing to add, that's what I expected.

You need this desperately!

https://tampax.com/en-us/shop-products

@blackhairedhero: It's fun watching you attempt to save face.

@fedor: Its fun watching you respond in short phrases because everytime you type a substantial amount you **** it up.

PS. Please change your user account name. It's a disgrace to have have a dunce like you disgrace the name of a once great MMA fighter.

@blackhairedhero: Whatever you need to think to help yourself feel better.

You really are dumb. Do you not understand what quotations are? You are literally not placing them correctly. Another failure on your part for not knowing how to write.

Of course it's all speculation. But they're more likely to speculate accurately than you

@neutrinoworks: They'll write 3-4 more articles speculating over the next year. Fact is nobody knows until we get 7nm Navi and Ryzen. Either way, the article you're referencing clearly states they are theorizing based of incomplete information. Have a good day, gio.

@Diddies:

"You are dumb as hell and so is your phone if it auto-corrects your text to not a real word."

I quoted the whole sentence for you. Did you get it this time? Now take a gander at this product of failure from our public school system. You write like you came out of a public school in Harlem.

@blackhairedhero: Ghetto? Harlem? You a bit racist?

@fedor: " You a bit racist?"

Jesus Christ please click on that link you sent me.

To answer your poorly worded question no. It's just a well known shitty public school system.

@blackhairedhero: "You write like you came out of a public school in Harlem."

"You sound like a Ghetto Hamster."

Seems like a racist to me. Maybe we should ask the mods?

@fedor: Ask the mods. You have been having a meltdown since the beginning of the thread. We can also ask the mods about your constant barrage of personal insults.

@blackhairedhero: Sure, enjoy your ban, racist loser.

@fedor: I literally sent you crying to the mods!

@blackhairedhero: I dont like racists.

That Radeon VII is actually AMD's M160 accelerator that's cut down. They are just creating an avenue to reduce wasted wafers that didn't make the cut to become the M160.

@fedor: Are you saying the Ghetto is associated with one race?

I don't like triggered babies.

@blackhairedhero: You've been triggered from the start. Thats why you came in calling people morons with that poorly put together post. Now you're showing your true racist self. I told you to quit while you were behind.

@fedor: I was never behind.

I made a simple comment about your shitty Tflop prediction and you got enraged. Then you bashed my sentence structure on a post only to butcher multiple post afterwords. Now you are grasping for straws calling me a racist. Meanwhile you couldn't post one valid argument to back up any of your predictions you just kept hurling personal insults at me, which in all honesty I found funny. You couldn't even stay on topic. Regardless you are not the forum police. I will respond when I feel like it.

Now you go cry to the mods.

@blackhairedhero: Its amazing how triggered you are. I know your scared about your impending ban, just relax. I make weak minded racists like yourself upset all the time.

Flagging me won't save you either.

@blackhairedhero: still correct. Lol

@fedor: lol I don't get scared about bans you pathetic loser. It's a gaming forum how pathetic are you? Is this all you have?

You are a bumbling moron and if I do get banned I'll pop on an alt and have a blast. You can't even come up with your own insults you baby. Bans here are an absolute joke. It's as simple as making a new email.

OHHHH NO! LOL

@blackhairedhero: Good.

From https://wccftech.com/exclusive-first-amd-navi-gpu-will-have-40-cus-and-is-codenamed-navi-12/

NAVI 12 has 40 CU with unknown chip size, but X1X GPU is already 44 CU with 280 mm2 on 16 nm process node.

Extra transistor budget with 7nm could go towards upgraded raster engines/ROPS, deep learning instruction set(e.g. ray trace de-noise), and BVH search tree accelerator (e.g. ray trace)

VII has other workstation hardware features which largely useless for consoles consoles like half rate FP64 and high bandwidth cache lower/slower/higher latency than L2 cache.

VII consumed extra 700 million transistors over Vega 64's 12.5 billion transistors for octa rate INT4 and half rate FP64.

Expensive part with VII is the four stack HBM v2. AMD could develop cheaper VII 56 at 1800Mhz with two stack HBM v2 512 GB/s 8GB model. It would be faster than Vega 56 at 1710 Mhz which beaten Strix Vega 64

For 7nm estimate, let's use AMD's conservative 70% shrink of 14nm Vega 10 (includes useless high bandwidth cache),

1. X1X's 44 CU GPU with 280 mm2 size yields 196 mm2, hence 64 CU version yields 285 mm2, hence it's 9.6 TFLOPS or 60 CU active yields 9.0 TFLOPS

Looking at 1.5X TFLOPS gain instead over X1X at the same clock speed.

X1X GPU has extra support for FP10.

2. Upgrade to 64 ROPS with 8 MB L2 cache would more than double effectiveness over X1X's 32 ROPS with 2MB render cache and TMU's 2MB L2 cache.

Looking at +2.0X ROPS gain over X1X at the same clock speed

AMD should have mastered 64 ROPS with 256 bit DDR6-16000 (352 GB/s) or 384 bit DDR6-12000 (576 GB/s). RTX 2060 uses 14000 rated GDD6.

With the same 384 bit bus as X1X's but with GDDR6-12000 substitution, yields 1.77X gain.

That's due to Pascal's problem with integer instructions.

https://devblogs.nvidia.com/nvidia-turing-architecture-in-depth/

Turing introduces a new processor architecture, the Turing SM, that delivers a dramatic boost in shading efficiency, achieving 50% improvement in delivered performance per CUDA Core compared to the Pascal generation. These improvements are enabled by two key architectural changes. First, the Turing SM adds a new independent integer datapath that can execute instructions concurrently with the floating-point math datapath. In previous generations, executing these instructions would have blocked floating-point instructions from issuing.

RTX 2080 has 545 mm2 chip size on 12 nm which is equivalent to 726 mm2 on 16 nm

Vega can mix integer and floating formats within a warp hence floating point instruction issues are not being blocked.

VII has 13.8 TFLOPS FP32 vs RTX 2080 has 10 TFLOPS FP32, hence VII has excess TFLOPS.

@loco145 said:

it took AMD $700, 300W and 7nm to reach what Nvidia did in 16nm 3 years ago. Is rather disappointing.

GTX 1080 Ti has 6 raster engine and 88 ROPS superiority up to 1800 Mhz (floats around 1700 Mhz) while Vega 64's four raster engines and 64 ROPS at 1536Mhz. This raster bottleneck is not an issue with cryptocurrency workloads for Vega 64 hence it rivals GTX 1080 Ti

VII is using 1800Mhz clock speed to close rasterization gap against GTX 1080 Ti, hence fastest 64 ROPS Turing card is RTX 2080.

@ronvalencia: First, yes. As your quote states, those operations in Turing independent from the Cuda cores which are now free! What GCN has is that a single compute unit can do both, floating point and integer, while with pascal and Maxwell, one type of operation would block the other on the whole SM. This is no longer an issue with Pascal and furthermore, the the full cuda cores are free of doing the tensor operations. What you are implying, that GCN can do integer operations for free (at the same time as a float operation on the same shader) is simply not true, but is the case with Pascal since is has significant silicon budget dedicated to it!

Your memory calculations fail to take into account that the bus is shared between the GPU and CPU in a SoC. Memory bandwidth wont be the strong point of the system, just like the PS4 had problems with AF, you can't hope to bail AMD architecture compared to grossly overclocking the memory on a card (yield issues anyone?) Also, you ignore that the GPU (at 7nm!) is already $700, so you are proposing to increase the number of rasters on the chip? So, you have an architecture that is bandwith memory starved, (the Vega 56 video you like to post so much) and you want to fix it by adding even more compute units that will require even more memory bandwith? The modded Vega 56 consumes over 400W btw.