Home
Creators
Search Recent Missing Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

Discussing Nvidia's Earnings, GPUs in 2023, and Zen 5 w/ AdoredTV (Patreon)

Published:
2023-05-27 02:36:57
Imported:
2026-06
Tags:

Content

Broken Silicon's next guest will be Jim from AdoredTV! We will be heavily discussing recent earnings from Intel, Nvidia, and AMD - and also the latest developments at Computex! (Feel free to submit questions right after Nvidia's Keynote!)

Everything from RDNA 4, Blackwell, Zen 5, Meteor Lake, Arrow Lake, consoles, and upcoming AMD APUs are all fair game. This episode's conversation will be heavily steered by your submissions below!

You have until Sunday at Midnight US Central Time (Right after Nvidia's Computex Keynote) to submit below...but please submit within 24 hours if it's not related to Computex! :)

Be concise, use good grammar, and be respectful to have your questions considered.

AdoredTV YouTube Channel: https://www.youtube.com/@adoredtv

Last Episode Jim was on: https://youtu.be/uKCaQOM6rUs

Comments

Roy Ackerman

Hi Tom and Jim, got any insight about the X86S and how is it going to impact the desktop market? Cheers from Israel, love your content keep up the great work! Roy

Kinihun25

Hey Tom and Jim do you see AMD and Nvidia making cards that are boring and another that is literally almost no improvement over last Gen, for Blackwell and RDNA4? I just get the feeling that the 7600 should have been the 7500XT and cost $199 and the 4060ti Should have been a 4060 and cost $299. I'm in the Enthusiast tier so this doesn't affect me but I have alot of friends that are midrange to low end. I know RDNA2 is a good option but that stock won't last forever. I just have a bad feeling for the future of pc gaming for people that want to game in the midrange/low end. I just hope this trend doesn't continue.

Falto

Looking at Nvidia current lineup, it's all a mess! Do you guys think they will have Lovelace refresh? Or we are not getting anything until blackwell? And speaking of which what do you guys expect from that architecture performance vram and price wise?

Shaun Ashment

How do you see AMD competing with NVIDIA in the commercial space? AMD is betting on an open source software solution while NVIDIA has their closed and currently much superior CUDA. Who do you think the commercial players will use? Will some take a chance on AMD so as not to get locked into one player? Can the biggest players just create their own GPU or equivalent?

Trogdor

Hi Tom and Jim, why do you think there is such a large discrepancy between the recent success and overall mind share of Ryzen and Radeon? Is it more to do with the recent performance of their respective competitors?

AI Generated Fish

Jim, your arrow lake leak was probably the one that got me the most excited... Until Tom dropped the strix halo leak. I'm interested to hear your take on how arrow lake is shaping up to take on what should be a really formidable competitor in strix halo.

Benjamin Oechsli (ex-lee)

Hey Tom and Jim, what great timing for this guest appearance! Jim releases an excellent video on the MI300 family (and the brewing AI competition in general), and a mere day later, Tom releases a video on Nvidia's financials, and how their AI customers are keeping them up right now. I want to hear a meeting of the minds: You only really need GPUs to *train* the AI, with CPUs able to handle the input/output. As these companies buy piles and piles of GPUs for training, is AI a bubble already preparing to burst? And will AMD's all-in-one offerings mean they come out on top of it? Thanks! Hope Jim's hay fever feels better soon.

Prior is the theme 999

Hello Tom and Jim, NVidia is betting on Ai going forward and they don't seem to care that gaming revenue is down tremendously. Do you think this bet will pay off? Will Ai "go to the moon" unlike the bet they made on crypto?

QuickJumper

Hi Tom and Jim concerning GPU strategy. Tom said that Nvidia basically doesn't care if AMD makes better gaming chiplets because they will just become AI company anyways. But is it possible to create most powerful ML chips without chiplet technology? Because it seems that Nvidia will still have to put lots of energy into chiplets.

CRCV

Hi Tom and Jim. The new console generation is now almost three years old. Usually, it would means that we would be close to the half-life of that gen. However, it is also the only console gen which has launched as the world almost stopped. Do you think that this will increase the lifespan of that gen to "compensate" for the pandemic (for example a 8 year gen)? Or will Sony and Microsoft will see an early enough uplift in performance from AMD to maintain a 6-7 year gen?

CRCV

CPU questions: There has been news reports recently of cooperation between both AMD and Arm, and Intel and Arm. What is the vision here? Blurring the line between x86 and arm? Or trying to come back to mobile market (phones, tablet, etc)?

Minikoopa

Initially, I dreaded the impact of the AI boom on consumer GPUs, envisioning a scarcity. Surprisingly, the outcome has been disappointing: exorbitantly priced gaming GPUs of subpar quality due to the overwhelming demand and profitability in enterprise markets. Considering the absence of compelling incentives to allocate wafer supply towards low-margin products, I wonder if this trend will persist. Frankly, it appears evident that Nvidia has completely lost interest in the gaming sector. Nevertheless, I appreciate your efforts and dedication.

TechnoLadz

Hey Tom and Jim. With Lovelace being priced severely poor, could we see Blackwell priced better. Typically; when we have a very minor price to performance increase (Fermi 2, Maxwell 2, Turing, etc), it’s usually followed up with a large performance increase with either a minor price increase or decrease (Ampere; “same price” for the tier as Turing; with the asterisk of the pandemic and mining, Pascal; slightly pricier than Maxwell for the tier. Kepler; slightly cheaper than Fermi 2 for the tier). However, with Lovelace being so much pricier than Ampere, could we see Blackwell return to a semi-sense of Normality; perhaps having the 5080 at $899? (Not great, but would at least be inflationary from 2014 prices of the GTX 980 being $649). Thanks for reading, and I await your response.

Dylan M

Hey Tom and Jim. You predicted Strix Halo might be good enough to kill low end GPUs on mobile. But it would also be nice to have on desktop, especially SFF builds. Despite not fitting on AM5 ,do you think AMD will try to put it on desktop? Could AMD make a new socket for it? Will this encourage AMD to replace AM5 sooner than expected? Could it remain a different socket like workstation? And how would you cool such a system? Thanks for reading my question.

Deadeyes

Hi Tom and Jim I keep hearing rumblings about new mid-gen console refreshes like we had with Xbone X and PS4 Pro, but to a large extent I don't see the point. With the previous refreshes 4k TVs had become standard and the console manufacturers needed to make something that would output a picture that wouldnt look like a blurry mess on them. With the current gen the only really meaningful performance goals I can come up with for them to chase are improved HFR or a dramatic increase in RT capabilities. But frankly I don't see where AMDs hardware has advanced enough to provide a really significant leap forward in either category without a price hike unless the next RDNA really steps up. Plus, I think having a lot of SKUs to develop for and validate contributes to games coming out broken or unfinished. Do you think we will actually see refreshes? If so how do you think they will differentiate themselves from current consoles? And do you think it would be a significant problem for developers?

Matheus Duque

Hi Tom and Jim! In the current market of gamer apathy and the AI bubble do you think AMD and Nvidia will actually care about making good products next gen? How do you guys feel about RDNA4? I remember Jim being quite bullish about it in the recent past, how's the sentiment now?

Alex Sloter

Hi Tom and Jim, I saw a headline the other day, "Nvidia Slashes Prices," And sure enough, Nvidia dropped prices on some Ampere cards. However, I don't see how that changes anything. The 3070 now costs the same amount as the 4060ti. Nvidia is willing to sell you 3070 performance for 3070 pricing across two different generations. Exciting. I am beginning to wonder if we are getting to a point where dropping prices will no longer entice consumers. By the time Nvidia drops prices on the midrange to a reasonable price/performance level, everyone who would have bought in that tier will have resolved to wait until Blackwell, purchased discounted RDNA2, or gotten Ampere from ebay. Thoughts?

KingHarkinian

Hello. On Jim's last visit, both of you said that you were "disappointed" that RDNA3 only chipletised the MCDs to one GCD, instead of doing full CCXes. AMD seems to be very lost right now with RDNA 3's performance. Do you still feel like they should've gone all the way?

KingHarkinian

AMD's 7600 is a tiny die made by a half-empty TSMC 6 factory. The entire point of the card is to be cheap. I found reviewers to be extremely petty about the launch price. Frankly, does it matter at all if it's $270 or $250, especially since the price will go down very quick? I believe the real debate should've been on the worth of the card itself, not its price. Should AMD (or anyone) be selling a 1080p 8Go VRAM card in 2023? Is it even worth buying any price? Is there a super-budget market that still will buy large volume 1080p hardware? In every way, value is now in the 1440p, 12Go VRAM minimum market. The 6700 xt is immensely more valuable than the 7600, even if the 7600 was $180 and the 6700 xt double that. "Should the card have been made at all?" is the real debate in my eyes. Thoughts on that?

Reginald Aryee

The topic of yields is important, as we all know thanks to Jim, but the specific topic of yields after the packaging process has been brushed over by most of the tech media. Considering Ponte Vecchio has 47 chipl---, sorry tiles, high yields after the packaging process must be more difficult to achieve compared to Zen configurations and probably significantly more costly too. How good or bad are Intel's yields for this kind of chip? Do you happen to know ballpark packing costs for a monolithic, chiplet, and chiplet monstrosity might be? Thanks. Images for the stream: https://www.tomshardware.com/news/intel-xe-hpc-ponte-vecchio-examined

Dave Scholze

Alright guys, how's it going? APU: What Performance could we expect from a Zen 5 chip with integrated RDNA 4 graphics? Will we see a better scalability with system memory to increase available (video) memory for the integrated graphics instead of 4 or 8GB only? CPU: With Zen 5 being kept at a maximum of 16 Cores for consumers, when will we be seeing 32 Cores for the mainstream? Thanks guys and take care! :D

Robert S Barnes

In your recent MI300 video you hinted that while chiplets let AMD throw more transistors at HPC and AI problems they are still behind Nvidia in software. At last years Financial Analyst Day Victor Peng made an under reported announcement which may have been the most important of the day. It was the architectures for the AMD Unified AI Stack 1.0 and 2.0. Have you heard anything about this effort and do you think AMD can close this software gap quickly enough to take advantage of the large language model inferencing driven AI boom? Slides 23-25 in Victor Peng's slide deck: https://d1io3yog0oux5.cloudfront.net/_decad72055e831bd2560ebeaeffa399e/amd/db/812/6836/presentation/FAD+2022_Victor_Peng_Final+Post.pdf

Ryzza5

Agree 100%, Until the majority of games fully utilise the PS5 potential I see no point in a mid-stage upgrade from Sony. I don't know about Xbox and frankly don't care as their main games are either MIA or also available on PC.

Ryzza5

Hi Tom & Jim. In your view what hypothetical stars have to align for Nvidia graphics card price to performance ratios to improve enough that most gamers would be happy or impelled to buy from them again? Whether they be internal or external factors.

【ｓｃｅｎｔｗａｖｅ】

When is N32 going to show, and are there going to be any more shuffling to cover the range between 7900XT and 7600 in the meantime?

【ｓｃｅｎｔｗａｖｅ】

Given the expected shrink of PHY+controller vs logic/SRAM/etc, what would a successor chip to N33 look like on N5P, what is AMD's next entry-level chip? Would it be monolithic or MCM, how many GDDR channels, any stacked components, etc? What do you think the IC module density will be for GDDR6, GDDR6W, or any other chosen memory subsystem? It kinda seems like everyone needs more density per IC even in this gen. And Infinity Link is potentially an interesting optimization, an infinity link PHY is much smaller than a full GDDR PHY, and thus the opportunity cost for disabling one is less etc. Maybe we see more 6+ PHY dies in the future. But monolithic is quite energy-efficient, and cost-efficient at smaller sizes due to lack of packaging. I'm curious how it would all shake out. PCB fanout is a cost optimization but increases energy vs interposer too.

Chris Rijk

Hi Tom, thank you for the fascinating video on Friday discussing initial 4060 Ti and 7600 sales. One of your sources in that Friday video suggested that sales could pick up over the weekend, so I was wondering if you have any further updates. Also, can you say anything about overall retail sales of GPUs? The data reported by TechEpiphany for sales in Germany suggests that AMD and Nvidia have been neck and neck for months and I was wondering how things compare in the US or rest of the world.

Chris Rijk

With the initial sales of the 4060 Ti and 7600, did we just witness a fundamental shift in the GPU market? Might the 4060 Ti 16GB and the standard 4060 also flop? If they do flop, would you expect Nvidia to just shrug it off? After all, they’ve done nothing much since the 4080 flopped half a year ago.

Chris Rijk

Hi Tom, in your video on Friday you said that Nvidia was diverting wafer allocation away from consumer graphics to high-end commercial parts. This makes sense given the poor sales of the former and the rampaging demand for AI. Do you guys expect to see Nvidia shift chip R&D resources as well? As well as their huge stockpiles, maybe this would explain why Blackwell seems to be delayed. Might Nvidia put minimum effort going forwards into their consumer GPUs?

Chris Rijk

I believe that most of Nvidia’s volume is in laptops and AMD doesn’t seem to be making much inroads here. However, with Meteor Lake causing OEMs to stop using lower-end discrete GPUs, followed by Strix Point, Strix Halo and Arrow Lake, do you guys expect to see a big change in market share in the next two years for low-end gaming laptops? Would you expect Nvidia to even do 106 and 107 dies with Blackwell?

SumRndmPenguin

Hi Tom and Jim. I know AMD is making great strides in their integrated graphics, but do you think Intel's integrated graphics will ever reach the same levels? I doubt they will due to AMD having a more competent graphics division but I'd like to hear your thoughts.

Bran Muffin

Do you think RTX 5090 will need PCIe 5.0 or will 4.0 likely be enough?