News AMD and Intel CPU Market Share Report: Recovery Looms on the Horizon

[Admin]

We take a look at the Mercury Research CPU market report for the first quarter of 2023.

AMD and Intel CPU Market Share Report: Recovery Looms on the Horizon : Read more

 

[aalkjsdflkj]

I'm curious about how AMD can have an overall x86 market share higher than all of the component market shares (mobile, desktop, server). The first graph makes it look like mobile, desktop, and server market shares are all between 15-20%, but the overall x86 looks like it's about 32%.
I tried to look at the Mercury website linked in the article, but didn't find the relevant information.
Also, I am color blind so it's possible I'm seeing the colors incorrectly. If anyone can explain what I'm missing I'd love to hear where I'm going wrong in looking at this chart. Perhaps there's another segment (maybe gaming consoles with the PS5/Xbox?) that isn't shown where AMD makes up the difference? If so why isn't that segment shown?

 

[TerryLaze]

Perhaps there's another segment (maybe gaming consoles with the PS5/Xbox?) that isn't shown where AMD makes up the difference? If so why isn't that segment shown?

From their latest quarterly report:
The Data Center segment primarily includes server microprocessors (CPUs) and graphics processing units (GPUs), data processing units (DPUs), Field Programmable Gate Arrays (FPGAs) and Adaptive System-on-Chip (SoC) products for data centers.
The Client segment primarily includes CPUs, accelerated processing units that integrate microprocessors and GPUs (APUs), and chipsets for desktop and notebook personal computers.
The Gaming segment primarily includes discrete GPUs, semi-custom SoC products and development services.
The Embedded segment primarily includes embedded CPUs and GPUs, FPGAs, and Adaptive SoC products.

There are another two: Gaming and embedded that would include consoles(semi-custom) but also steamdeck and the likes(embedded-soc) .

Mercury is a PC parts analyzer so maybe they don't have the connections or just the know-how to analyze the console market.

 

[PaulAlcorn]

I'm curious about how AMD can have an overall x86 market share higher than all of the component market shares (mobile, desktop, server). The first graph makes it look like mobile, desktop, and server market shares are all between 15-20%, but the overall x86 looks like it's about 32%.
I tried to look at the Mercury website linked in the article, but didn't find the relevant information.
Also, I am color blind so it's possible I'm seeing the colors incorrectly. If anyone can explain what I'm missing I'd love to hear where I'm going wrong in looking at this chart. Perhaps there's another segment (maybe gaming consoles with the PS5/Xbox?) that isn't shown where AMD makes up the difference? If so why isn't that segment shown?

Hi, thanks for reading. The section at the bottom of the article with the 'overall x86 market' breakdown has the explanation underneath. It reads:

"Whereas the other segments above exclude IoT and semi-custom units (like AMD's game console business), this accounting of the overall x86 market also includes those products."

You'll also see that there are tables at the bottom for each segment, with the full raw numbers presented.

 

[aalkjsdflkj]

Thanks!

 

[ikjadoon]

That's actually a surprisingly resilient unit share considering it has been six years since AMD launched its incredible comeback with its first-gen Ryzen PC chips back in 2017.​

I frequently find these reports kind of meaningless because it's all MARKET share, not usage share. Market share just counts unit sales.

For servers, sometimes procurement buys 1x new CPU-A to replace 3x old CPU-Bs. But for market share, CPU-B had a higher market share than CPU-A, even though 3x CPU-B is no longer being used. It was sold, so it gets counted.

Ideally to answer how "resilient" Intel (or AMD or Arm or anyone) is, we'd need quarterly usage, not quarterly sales. Sales #s are still neat, but they can be misleading.

 

[suryasans]

This report explains why AMD's consumer marketshare is decreasing. The DDR5-DDR4 RAM price crossover has not been reached yet while DDR4 is unsupported with AMD's new AM5 socket. Almost new entry level DIY consumers will buy PC based on Intel Core i5 13400 with cheaper Dual 16 GB DDR4-3200 SDRAM than new DDR5 based AMD's AM5 based Ryzen.

 

[msroadkill612]

This report explains why AMD's consumer marketshare is decreasing because DDR4 has beeb unsupported with its new AM5 socket. Most new entry level DIY consumer will buy PC baee on Intel Core i5 13400 with cheaper Dual 16 GB DDR4-3200 SDRAM than new DDR5 based AMD's AM5 based Ryzen.

DDR5 remedies one of the biggest bottlenecks to improved perf - ram bandwidth. The consensus is u would be a mug to get a new rig w/ ddr4 at current prices... if u were recycling a big investment in ddr4 from ur old rig maybe.

 

[msroadkill612]

""Intel is seeing increasing stability in the data center and PC markets."

Why cant they say anything sensible?

This could mean its flatlining at the lowest point.

 

[suryasans]

DDR5 remedies one of the biggest bottlenecks to improved perf - ram bandwidth. The consensus is u would be a mug to get a new rig w/ ddr4 at current prices... if u were recycling a big investment in ddr4 from ur old rig maybe.

Almost entry level DIY PC Buyers do not plan an upgrade to their new build PC rigs. With limited budget They prefer to buy NVME SSD with hard drive than fastest RAM available. AMD's wrong prediction on DDR4-DDR5 price crossover resulted with its marketshare dwindled. https://www.pcgamer.com/amd-ddr5-pricing/

 

[TerryLaze]

AMD's wrong prediction on DDR4-DDR5 price crossover resulted with its marketshare dwindled.

It's not a wrong prediction, they knew this, this is just a business decision, they couldn't afford to make both work so they focused on which will have a longer life time.
At least that's my opinion because I don't hink that lucy lu is that stupid.

 

[dk382]

""Intel is seeing increasing stability in the data center and PC markets."

Why cant they say anything sensible?

This could mean its flatlining at the lowest point.

I mean, just look at the lines. Intel still has around 82% market share for servers. I think they'll be quite happy if their market share "flatlines" at that point considering how aggressive AMD has been.

 

[msroadkill612]

Almost entry level DIY PC Buyers do not plan an upgrade to their new build PC rigs. With limited budget They prefer to buy NVME SSD with hard drive than fastest RAM available. AMD's wrong prediction on DDR4-DDR5 price crossover resulted with its marketshare dwindled. https://www.pcgamer.com/amd-ddr5-pricing/

For the budget constrained, AMDs market share in the bell weather DIY market has boomed - only its been in am4 ddr4 platforms.

If you want the best in gaming, its the 7800x3d, & for productivity, its the 7900 or the 7950, or at both, the 7950x3d.

its the coolest (hence cheaper cooling), most efficient (cheaper psu) & a lot more IO (24 vs 16 pcie5 lanes) & several gens of new supported cpus to come vs nil for intel

Intel can contrive a few apparent benchmark wins, but folks are opting for the balance amd offer, afaict on buildapc sites etc

 

[TerryLaze]

its the coolest (hence cheaper cooling),

A $20 cooler gives you the full warrantied performance of the 13900k ,I don't think that you can go much cheaper than that on cooling for anything...

Core i9-13900K Cooling Guide: Testing Intel's Flagship With Budget Air and Big AIOs

The Core i9-13900K can push even the most capable liquid coolers to their limits, but you don’t lose that much performance if you go with budget-priced air cooling.

www.tomshardware.com

Also with the same amount of cooling the 13900k is cooler than at least the 7950x at any power level, sadly no info on the 3xd they might be better.

A Lighter Touch: Exploring CPU Power Scaling On Core i9-13900K and Ryzen 9 7950X

www.anandtech.com

most efficient (cheaper psu)

Only if you plan to use your PC without an GPU, and even then you might win about 100W if you overclock the intel system to the limit compared to the stock ryzen, it will be about 200W for ryzen compared to 300-350 for overclocked intel.
This will be the same level of PSU for both, nobody is going to use a 200W PSU on an expensive CPU like that, even 450-500W would be a huge stretch.

You save a bit on the electricity bill but then you can also just set the power limit for the intel to the warrantied limit of 253W or even lower.
As the first link above shows the difference in performance between a $20 ( around 240W) and a $200 cooler (315W) is about 6%

 

[msroadkill612]

A $20 cooler gives you the full warrantied performance of the 13900k ,I don't think that you can go much cheaper than that on cooling for anything...

Core i9-13900K Cooling Guide: Testing Intel's Flagship With Budget Air and Big AIOs

The Core i9-13900K can push even the most capable liquid coolers to their limits, but you don’t lose that much performance if you go with budget-priced air cooling.

www.tomshardware.com

Also with the same amount of cooling the 13900k is cooler than at least the 7950x at any power level, sadly no info on the 3xd they might be better.

A Lighter Touch: Exploring CPU Power Scaling On Core i9-13900K and Ryzen 9 7950X

www.anandtech.com

Only if you plan to use your PC without an GPU, and even then you might win about 100W if you overclock the intel system to the limit compared to the stock ryzen, it will be about 200W for ryzen compared to 300-350 for overclocked intel.
This will be the same level of PSU for both, nobody is going to use a 200W PSU on an expensive CPU like that, even 450-500W would be a huge stretch.

You save a bit on the electricity bill but then you can also just set the power limit for the intel to the warrantied limit of 253W or even lower.
As the first link above shows the difference in performance between a $20 ( around 240W) and a $200 cooler (315W) is about 6%

ur right,
hotter, louder, less efficient & spikey is more fun to have on your desk & doesn't add hidden extra costs..

 

[TerryLaze]

ur right,
hotter, louder, less efficient & spikey is more fun to have on your desk & doesn't add hidden extra costs..

That always applies to any overclock, and it also always applies that you are never forced to run your CPU overclocked.

 

[msroadkill612]

That always applies to any overclock, and it also always applies that you are never forced to run your CPU overclocked.

Nor forced to not overclock ur AMD system to exceed their perf claims, vs underclock ur intel to give lie to their civilized real world perf claims.

IE they forget to say u need expensive CPU cooling, A/C, & ~1kw PSU in case of spikes. Adds up to dear & no fun, as i say.

Intel is the company who thinks it is kosher to hide an industrial scale multi kw cooler unit behind the stage's curtains during an alleged demo of an upcoming product.

 

[TerryLaze]

Nor forced to not overclock ur AMD system to exceed their perf claims, vs underclock ur intel to give lie to their civilized real world perf claims.

IE they forget to say u need expensive CPU cooling, A/C, & ~1kw PSU in case of spikes. Adds up to dear & no fun, as i say.

Intel is the company who thinks it is kosher to hide an industrial scale multi kw cooler unit behind the stage's curtains during an alleged demo of an upcoming product.

So how does any of that change the numbers from the links?!


The $20 cooler still keeps the 13900k at around 250W with the performance not changing enough for anybody to notice, and the CPU can't have any spikes above what the cooling allows, a CPU isn't a GPU, that is a problem with GPUs that can draw more than twice the power for a moment, CPUs can't do that because they will thermal throttle way too fast for the power to get much higher than what your cooling can provide.
Recent ryzen 7000 models excluded....and even then it's rare for it to happen.

 

[bit_user]

Also with the same amount of cooling the 13900k is cooler than at least the 7950x at any power level, sadly no info on the 3xd they might be better.

A Lighter Touch: Exploring CPU Power Scaling On Core i9-13900K and Ryzen 9 7950X

www.anandtech.com

You love to keep citing that, but die temps really don't matter much. In either Intel or AMD's case, better cooling leads to a little bit better boosting, but that's about it.

 

[TerryLaze]

You love to keep citing that, but die temps really don't matter much. In either Intel or AMD's case, better cooling leads to a little bit better boosting, but that's about it.

Yes, every time someone regurgitates the lies they hear about ryzen being cooler and/or easier to cool I will keep citing this.
You can cool the hottest intel CPU with a $20 air cooler while you need a high end cooler for any ryzen.

In either Intel or AMD's case, better cooling leads to a little bit better boosting, but that's about it.

Yeah and then you complain that I keep posting the same link, maybe read it sometime....
For intel better cooling means that it will boost better, the 7950x barely hits its advertised TDP with one of the best AIOs the review could get their hands on, there is no better boosting to have there unless you go exotic or you make your system less stable by undervolting.

 

[bit_user]

Yes, every time someone regurgitates the lies they hear about ryzen being cooler and/or easier to cool

That was about the 7950X3D, so it's not a lie. Your calling it lies is a lie.

You can cool the hottest intel CPU with a $20 air cooler while you need a high end cooler for any ryzen.

Not really. The same cooler will also unleash most of Ryzen's performance.

Yeah and then you complain that I keep posting the same link, maybe read it sometime....

I did, and there's nothing inconsistent with what I said about boosting a little better. As a matter of fact, Raptor Lake is more sensitive to cooling than Ryzen, performance-wise. You love to cite that $20 cooler, but it's the only $20 that cools so well. Nearly any other $20 cooler will throttle the i9-13900K more substantially.

 

[TerryLaze]

That was about the 7950X3D, so it's not a lie. Your calling it lies is a lie.

What numbers do you look at for the claim about the 7950x3d?!
Also the exact quote was about the 7950x as well:
"If you want the best in gaming, its the 7800x3d, & for productivity, its the 7900 or the 7950, or at both, the 7950x3d."
If you try to lie at least make sure that the truth isn't on the same page just a few posts above.

Not really. The same cooler will also unleash most of Ryzen's performance.

I would like to see that, is there any review that tried it?
The ryzen already hits 94 with this high end cooler and AMD has an history of CPUs just shutting down at a certain point instead of throttling.

You love to cite that $20 cooler, but it's the only $20 that cools so well. Nearly any other $20 cooler will throttle the i9-13900K more substantially.

Again what are you basing your opinions on here?
Do you have cooler benchmarks you can link that show that other coolers at the same price point are much worse?!