AMD Execution Thread [2024]

[Seanspeed]

So less consoles being made? Surely then console sales also down large percentages.

Sony's already been forecasting these declines:

Sony narrowly missed its revised down target for PlayStation 5 sales. The firm said that sales of its flagship console totalled 20.8 million in the fiscal year 2023.

That’s slightly lower than the revised 21 million unit target that Sony gave investors in February. Prior to that, the company had forecast that its PS5 console would sell 25 million units for the full year.

Sony expects even weaker sales of 18 million units of its PS5 in the year ending March 2025, a company executive said, according to Reuters.

https://www.cnbc.com/2024/05/14/sony-q4-and-full-year-2023-earnings.html

And Xbox would seemingly struggle to give away Xbox's for free at this point.

 

[Shifty Geezer]

PS5 isn't as huge a downturn as the 50% gaming revenue decline, but I guess XB would count for a lot. And if Radeon in as well with that lot, it's not simply the console market shrinking 50%.

 

[Erinyes]

Supposedly yes. Tom's Hardware claims same goes for Ryzen 5's,

Also it's a lot less messed up compared to what they had at Tech Day
View attachment 11794
(image from https://www.tomshardware.com/pc-com...to-typo-ryzen-7-9700x-ryzen-5-9600x-confirmed )

Then there's Dr. Ian Cutress

https://twitter.com/x/status/1816473099468197980

I would be really surprised if true. It's not as if it's the first time they're doing this, this has been the naming convention since like 5 years now (with R9 since Ryzen 3000 I mean, even longer if you take Ryzen 1000) and mass production would have started weeks ago. To be caught this late would be quite the miss.

I'd seen that and I genuinely thought Ian was joking there.

PS5 isn't as huge a downturn as the 50% gaming revenue decline, but I guess XB would count for a lot. And if Radeon in as well with that lot, it's not simply the console market shrinking 50%.

Keep in mind console revenue is most likely only royalty and not the full cost of the SoC. So a decline in sales of even both consoles wouldn't be as much of an overall decline to AMD's gaming segment revenue as a decline in gaming GPUs. This hints that gaming GPU revenue fell more than console revenue. Seems desktop GPU sales fell significantly. Notebook GPU sales would hardly have had an impact as AMD's market share is negligible in that segment anyway.

 

[DavidGraham]

Keep in mind console revenue is most likely only royalty and not the full cost of the SoC. So a decline in sales of even both consoles wouldn't be as much of an overall decline to AMD's gaming segment revenue as a decline in gaming GPUs

Well, there is nuance to that. In 2022 (a year where AMD had low Radeon market share), Sony accounted for 20% of revenue of the entirety of AMD (without Xilinix), that's Sony alone, not including Xbox (likely another 10%).

https://twitter.com/x/status/1630600116339081217

So consoles revenue was already huge for AMD, but not much profitable, meaning consoles give AMD big gaming revenue, but with marginal profits (~15%). For example, from 2021 to 2022 AMD had an extra 1.2 billion of gaming revenue from increased console sales but only 19 million in extra profits.

Some people analyzed their 2022 numbers and came to the conclusion that AMD's entire dGPU sales in the whole of 2022 is only around 1 billion$, with profits in the range of 100 mil to 200 mil.

AMD had 6.8 billion in total gaming revenue for the whole of 2022, consoles accounted for 5.7 billion (84%), dGPUs accounted for 1.1 billion (16%). Profits were 950 million for consoles + dGPUs combined (likely ~740 million for consoles and ~210 million for dGPUs).

2024 is a repeat of 2022 so far, with even lower Radeon market share numbers.

 

[ka_rf]

I would be really surprised if true. It's not as if it's the first time they're doing this, this has been the naming convention since like 5 years now (with R9 since Ryzen 3000 I mean, even longer if you take Ryzen 1000) and mass production would have started weeks ago. To be caught this late would be quite the miss.

I'd seen that and I genuinely thought Ian was joking there.

Keep in mind console revenue is most likely only royalty and not the full cost of the SoC. So a decline in sales of even both consoles wouldn't be as much of an overall decline to AMD's gaming segment revenue as a decline in gaming GPUs. This hints that gaming GPU revenue fell more than console revenue. Seems desktop GPU sales fell significantly. Notebook GPU sales would hardly have had an impact as AMD's market share is negligible in that segment anyway.

Radeon sales were up, console sales crashing are what caused the big drop in AMD's gaming revenue.

AMD sells the full chip to MS and Sony, it's not a royalty deal.

 

[Frenetic Pony]

I want a thin and light AI 370 laptop. I just want one that has a better keyboard and ssd than the Asus stuff

 

[Seanspeed]

Radeon sales were up

Source? We never get any good info on this.

 

[TopSpoiler]

Source? We never get any good info on this.

From their earnings conference call:

Now turning to our Gaming segment. Revenue declined 59% year-over-year to $648 million as semi-custom SoC sales declined in line with our projections. Semi-custom demand remains soft as we are now in the fifth year of the console cycle, and we expect sales to be lower in the second half of the year compared to the first half. In gaming graphics, revenue increased year-over-year, driven by improved sales of our Radeon 6000 and 7000 Series GPUs in the channel.

 

[DavidGraham]

From their earnings conference call:

It's a meaningless statement though, Radeon sales were very weak in Q2 2023, it amounted to only 10% market share, so if their Q2 2024 sales increased by 1% they would still call it improved sales, but it's meaningless and it's still weak sales. Their overall gaming revenue dropped so much that any claimed improved Radeon sales can't be that much to begin with.

Q2 2023 marketshare numbers are in, NVIDIA sets at 87%, AMD at 10% and Intel at 3%.

GPU Market Rebounds In Q2 2023: AMD, NVIDIA & Intel See Increased Shipments, Discrete GPU Up By 12.4%

JPR has posted KJQ2 2023 market analysis, showing increased GPU shipments for AMD, Intel, & NVIDIA thanks to market normalization.

wccftech.com

 

[Seanspeed]

Bit of a disaster, honestly. I knew it wasn't gonna be a great gaming uplift, but this was worse than expected. Even other workloads aren't improved all that much.

Also frustrating how many people are trying to defend it because of 'big efficiency improvements' even though all they did was lower the TDP, and too many people dont seem to know you could also run the 7700X at 65w. Actual architectural efficiency improvements seem fairly minimal.

 

[Kaotik]

There's something funky in Hardware Unboxeds CB24 numbers, supposedly his 9700X is drawing nearly 60 watts more than ours while scoring barely any better, while difference between their and our 7700X is just about 20W while they score the same. (so one could assume his system draws about 20W more than ours, which would leave 9700X drawing nearly 40 watts more than ours)
(it's in finnish https://www.io-tech.fi/artikkelit/testissa-amd-ryzen-7-9700x-zen-5/ )

 

[Subtlesnake]

Also frustrating how many people are trying to defend it because of 'big efficiency improvements' even though all they did was lower the TDP, and too many people dont seem to know you could also run the 7700X at 65w. Actual architectural efficiency improvements seem fairly minimal.

ComputerBase has it 9% faster than the 65W 7700 in multi threaded benchmarks. It's really lacking the faster clocks that Zen 4 benefitted from.

It's also interesting that it seems to do much better in single threaded benchmarks, even relative to the 7700X, which is only 100 MHz behind on boost clocks.

 

[DegustatoR]

It's also interesting that it seems to do much better in single threaded benchmarks, even relative to the 7700X, which is only 100 MHz behind on boost clocks.

Not that interesting as ST benchmarks aren't limited by the overall PL and you get the effects of clocks+architecture which in case of Zen 5 should provide at least 15% improvement over Zen 4 on the same clock.

 

[Subtlesnake]

Not that interesting as ST benchmarks aren't limited by the overall PL and you get the effects of clocks+architecture which in case of Zen 5 should provide at least 15% improvement over Zen 4 on the same clock.

The boost clock on the 9700X is only 1.8% higher than the boost clock on the 7700X though. And would we not expect a 15% IPC improvement to show in the multithreaded benchmarks vs the 7700, considering the 9700X clocks almost the same as the 7700?

 

[DegustatoR]

And would we not expect a 15% IPC improvement to show in the multithreaded benchmarks vs the 7700, considering the 9700X clocks almost the same as the 7700?

Ideal MT workloads are heavily power limited on the other hand so a 65W part will inevitably lose in these to a 125W (or was it 145?) one despite architectural improvements.
It is also where 9700X shows the most perf/watt gains though as it ends up at about the same performance while consuming way less power.

All in all I agree with GN that the balance of power/perf/price on 9700X specifically seems weird to a point where you wonder why would it even be launched like this? Maybe AMD knows something about the upcoming ARL counterpart?

 

[Subtlesnake]

Ideal MT workloads are heavily power limited on the other hand so a 65W part will inevitably lose in these to a 125W (or was it 145?) one despite architectural improvements.
It is also where 9700X shows the most perf/watt gains though as it ends up at about the same performance while consuming way less power.

Right, but the 7700 is a 65W part too and the TPU review has the 9700X clocked higher until 14 threads are in use and with all 16 threads in use clocks are less than 1% higher on the 7700. So I don't see why we shouldn't expect the 9700X to beat the 7700 by ~15% in multi threaded benchmarks if IPC has really increased by that much.

 

[DegustatoR]

So I don't see why we shouldn't expect the 9700X to beat the 7700 by ~15% in multi threaded benchmarks if IPC has really increased by that much.

It's an average. Means that some workloads will get 0% while some may get 25%. Without knowing which one got what and what is being used for MT benchmarks it's hard to say if the result is out of line or not.

 

[homerdog]

Why did they limit TDP to 65W?

 

[Frenetic Pony]

Bit of a disaster, honestly. I knew it wasn't gonna be a great gaming uplift, but this was worse than expected. Even other workloads aren't improved all that much.

Also frustrating how many people are trying to defend it because of 'big efficiency improvements' even though all they did was lower the TDP, and too many people dont seem to know you could also run the 7700X at 65w. Actual architectural efficiency improvements seem fairly minimal.

It's a fantastic architecture, for laptops and datacenter.

AMD said it wasn't for desktop gaming, fair enough, so all these youtuber reviews are whinging about something everyone was already warned about.

But it's also not for gaming because AMD's priority of selling desktop CPUs to gamers is seconhand at best. Gamers get the X3d series months down the line, and only when inventory is free from the far more profitable laptop/datacenter market is even available. And datacenter is on 3nm while consumer is on 4nm now, meaning more inventory for gamers, but only after laptops (why Strix Point launched first). Thus gamers are third tier customers because the profit margin x revenue total just isn't there. That AMD currently isn't blowing away their tier 3 customers with the newest launch isn't really a big concern for them finance wise, especially with the lucky break of Intel screwing over the same customer segment at the same time.

 

[Kaotik]

Why did they limit TDP to 65W?

"Because it's so efficient". Only real reasons I can see are
- Planning to launch 9800X soon with 105W
- Trying to make 9900X look better
If you don't use AVX512 there isn't really much point on getting 9000, we'll see if X3D changes things.