[JPR] The balance of power in gaming

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

AtenRa

Lifer
Feb 2, 2009
14,000
3,357
136
AMD lost .5% total market share in q3, and probably a substantial percentage of the dGPU market to Nvidia.

Ehm have a look again at the numbers,

AMD gained Total marketshare from 12.3% in Q2 to 12.6% in Q3 while NVIDIA remained at the same 16.1%.

AMD gained 2.7% in Desktop marketshare

AMD only lost market share in Mobile.
d9175582_d885af5462868ad04d50c77c69de3d79.png



Even ignoring the facts, a shift of a few percentages for a quarter is not going to be reflected in installed user base surveys like steam. The math doesn't work that way.


The fact is that AMD gained marketshare in Q3 in Desktop were the vast majority of Steams Hardware Survey DX-12 GPUs are counted for, but its the NVIDIA DX-12 GPUs that are increasing there share in Steam Hardware Survey.

That is why Steam Hardware Survey doesnt reflect Quartelly GPU shipments and should not be used for this.
 

Pariah

Elite Member
Apr 16, 2000
7,357
20
81
Guess it depends who you ask.

http://wccftech.com/nvidia-amd-intel-gpu-market-share-q3-2016/
NVIDIA Gains Graphics Market Share in Q3 2016, Discrete GPU Shipments Up 39.8%, AMD Gains 34.7% – Intel and AMD Witness Slight Market Share Loss

Q3-2016-GPU-Market-Share.jpg

If Nvidia buried AMD as it appears they did last quarter. Then I'm not sure why you would question NVidia gaining in the DX12 subsection of the overall market. There were likely far more users that dumped AMD for a high end Nvidia card in q3 than there were Nvidia users that switched to a 480 or lower AMD card. Again, you're ignoring the difference between quarterly results and installed userbase that Steam is showing.
 

AtenRa

Lifer
Feb 2, 2009
14,000
3,357
136
The difference is that Mercury Research is only counting PC shipments (Desktop and Laptop), the ones we are talking about.
JPR takes the entire GPU market including Professional and HPC shipments.

So again, AMD didnt lose marketshare in Desktop which is a huge part of the Steam Hardware Survey. Mercury Research numbers are the ones to use for what we compare here.
 
  • Like
Reactions: USER8000

tential

Diamond Member
May 13, 2008
7,355
642
121
That's odd. Everywhere I read I hear about Polaris being amazing and boosting amds share. Now they lost share?

Confused a bit. This fits my personal narrative of the rx 480 being a lackluster gpu (being polite) and not having anything of merit to warrant amd gaining gpu share because of it.

But hey, I could be wrong. So did amd gain or lose market share after Polaris release? Or is this data delayed and doesn't include Polaris.

Edit : I'm so happy there is no way I'll embarrass myself and have to buy Polaris the way I have to buy a fury x now. I know fury x is a bad gpu but I have no option.

Also, amd prices dropped so much at the high end with fury at ~260 and fury x at 350 that Polaris rx480 8gb is now regularly below $200. Jesus amd pricing is always insanely weird. I love it, lots of room for exploitation, but still weird.
 

bystander36

Diamond Member
Apr 1, 2013
5,154
132
106
The difference is that Mercury Research is only counting PC shipments (Desktop and Laptop), the ones we are talking about.
JPR takes the entire GPU market including Professional and HPC shipments.

So again, AMD didnt lose marketshare in Desktop which is a huge part of the Steam Hardware Survey. Mercury Research numbers are the ones to use for what we compare here.
Steam surveys are going to show much lower changes, because it includes what is currently used and new shipments. That Mercury Research only counts new GPU's, which is a much smaller total. And let's not forget, the Mercury Research also includes a segment not included in the Steam reports.
 
Aug 11, 2008
10,451
642
126
AMD for a brief period at least had the midrange segment to themselves and should have gained some market share. Now however, they are competitive in the midrange, outclassed in the low end and of course not even competing in the high end. Despite all the hype for polaris, I can't think of a time when they were overall less competitive in the dgpu market than they are now. I am honestly amazed that their stock has not fallen off a cliff.
 

sirmo

Golden Member
Oct 10, 2011
1,012
384
136
The difference is that Mercury Research is only counting PC shipments (Desktop and Laptop), the ones we are talking about.
JPR takes the entire GPU market including Professional and HPC shipments.

So again, AMD didnt lose marketshare in Desktop which is a huge part of the Steam Hardware Survey. Mercury Research numbers are the ones to use for what we compare here.
Steam hardware market share requires the user to participate in it.. and I rarely ever see it pop up for me. I don't know how reliable it is at all. Perhaps it gets triggered when you first install the client on a new build or something, and since AMD isn't playing in the enthusiast market most people installing rx480s or rx470s are upgrading not building new PCs.. so it's a possible reason for why AMD is being under represented.

Also reading people question JPR numbers in this thread is so silly. Like 30% of desktop and 100% of the console market.. why wouldn't AMD be in the lead? It's common sense. AMD finally has the ecosystem, something they never had in the past. Developers can no longer ignore AMD, which is a nice change.

I must say I am really impressed with the business acumen of this new AMD leadership.
 

Pariah

Elite Member
Apr 16, 2000
7,357
20
81
AMD for a brief period at least had the midrange segment to themselves and should have gained some market share. Now however, they are competitive in the midrange, outclassed in the low end and of course not even competing in the high end. Despite all the hype for polaris, I can't think of a time when they were overall less competitive in the dgpu market than they are now. I am honestly amazed that their stock has not fallen off a cliff.

Can only guess investors are hoping for a miracle with Zen and Vega. If Vega disappoints (or more likely both), expect the stock to tank and AMD ceasing to exist as the company we have known since they acquired ATI. They've already been irrelevant in the CPU space for over a decade, and Zen is not going to be direct competition for Kaby Lake; Vega flopping will mark the end of their ability to compete in the GPU market and the end of any realistic hope of AMD's computing and graphics division of ever being profitable again. AMD's enterprise/embedded/custom division is profitable and may be all that is left of consequence by the end of next year.
 

AtenRa

Lifer
Feb 2, 2009
14,000
3,357
136
That's odd. Everywhere I read I hear about Polaris being amazing and boosting amds share. Now they lost share?

Confused a bit. This fits my personal narrative of the rx 480 being a lackluster gpu (being polite) and not having anything of merit to warrant amd gaining gpu share because of it.

But hey, I could be wrong. So did amd gain or lose market share after Polaris release? Or is this data delayed and doesn't include Polaris.

AMD outsold NVIDIA in Desktop volume shipments in Q3 2016, that is documented in the Mercury Research table above.

But in the overall GPU market that consist of both PC (Desktop+Laptops) and Professional/HPC , NVIDIA shipments were higher and thus why the JPR numbers talk about NVIDIA gaining share.

Since we are concerned ONLY with the desktop market here, I will say that Polaris have sold very well till now. The only problem is that because of the miners, fewer Gamers bought Polaris cards in the first 1-2 months and that is displayed in the Steam Hardware Survey.

That is why im debating that we shouldnt take Steam Hardware Survey to illustrate Global GPU shipments.
 

bystander36

Diamond Member
Apr 1, 2013
5,154
132
106
AMD outsold NVIDIA in Desktop volume shipments in Q3 2016, that is documented in the Mercury Research table above.

But in the overall GPU market that consist of both PC (Desktop+Laptops) and Professional/HPC , NVIDIA shipments were higher and thus why the JPR numbers talk about NVIDIA gaining share.

Since we are concerned ONLY with the desktop market here, I will say that Polaris have sold very well till now. The only problem is that because of the miners, fewer Gamers bought Polaris cards in the first 1-2 months and that is displayed in the Steam Hardware Survey.

That is why im debating that we shouldnt take Steam Hardware Survey to illustrate Global GPU shipments.
If you count IGP's, but not in discrete sales. Otherwise you are misreading what was reported.
 

AtenRa

Lifer
Feb 2, 2009
14,000
3,357
136
If you count IGP's, but not in discrete sales. Otherwise you are misreading what was reported.

What report are you talking about ??? Because Mercury Research report, which im using, only takes Desktop/Laptop Discrete GPUs and that is what im using.
 

Samwell

Senior member
May 10, 2015
225
47
101
What report are you talking about ??? Because Mercury Research report, which im using, only takes Desktop/Laptop Discrete GPUs and that is what im using.

Mercury Research is showing average market share over 4 last quarters. This is of course increasing, because 29,1% in Q3 is still way better than the 20-22% in Q4 15 and Q1 16. If you want to see real market share in Q3 than you need to take JPR numbers, which show that amd slightly lost market share compared to Q2.

Also i don't know how you came to the conclusion, that Amd outsold Nvidia. Desktop shipments were alright for not having gpus in higher segments, but in numbers they sell 1 GPU für ~2,5 which nvidia sells. Also nvidia increased their sales in Q3 more compared to amd .
AMD’s .. Desktop discrete GPUs increased 34.7% from last quarter, and notebook discrete shipments increased 23.0%.
Nvidia’s desktop discrete GPU shipments were up 39.8% from last quarter; and the company’s notebook discrete GPU shipments increased 38.7%, and total PC graphics shipments increased 39.3% from last quarter.
https://www.jonpeddie.com/publications/market_watch/
 
Last edited:

Pariah

Elite Member
Apr 16, 2000
7,357
20
81
If you count IGP's, but not in discrete sales. Otherwise you are misreading what was reported.

Even including IGP's, Nvidia is leading 16.1 to 13.1 according to his own chart. Limiting to descrete, it's still the same roughly 80% to 20% lead Nvidia has had since the release of Maxwell. Maybe he has another chart he hasn't shown the rest of us. I don't know.
 

USER8000

Golden Member
Jun 23, 2012
1,542
780
136
All the E-PEEN bickering and hoping one company goes bankrupt over another one here has missed one important thing - Q3 2016 is the best quarter for dGPU companies since 2014.

That means both Polaris and Pascal have been successful in getting people to upgrade more of their systems,despite the PC market shrinking yet again.

Marketshare has been gained from Intel mostly - this is why Intel has dropped overall in share. That shows more people have upgraded their systems with a dGPU from integrated graphics.
63b76f8b-b80a-48ab-80a4-d748ae9a2945.PNG


Look at the segments which rebounded the most even from one year ago - performance and mainstream.
 
Last edited:

bystander36

Diamond Member
Apr 1, 2013
5,154
132
106
What report are you talking about ??? Because Mercury Research report, which im using, only takes Desktop/Laptop Discrete GPUs and that is what im using.
Then you are misreading the results. Just because AMD increased sales, and Nvidia decreased sales over 3rd quarter does not show AMD selling more than Nvidia. It shows AMD selling more than they did the previous quarter. Those are not numbers showing the percentage of AMD and Nvidia users. That is a chart showing the sales numbers for those quarters.
 

AtenRa

Lifer
Feb 2, 2009
14,000
3,357
136
Ok my bad i didnt word it as i would like, "outsold" was not the correct word to use.

It is obvious that NVIDIA did shipped more volume than AMD in Q3 for Desktops (74.5% vs 25.5% Mercury), but the think is that AMD managed to ship more volume than the previous Quarters in a declining market. That made its market share to grow vs NVIDIA. That gained didnt take place in the Steam Hardware Survey (last 3 months) for many reasons and this is why im saying that Steam Hardware Survey shouldn't be used to illustrate global desktop volume shipments.
 

Phynaz

Lifer
Mar 13, 2006
10,140
819
126
Ok my bad i didnt word it as i would like, "outsold" was not the correct word to use.

It is obvious that NVIDIA did shipped more volume than AMD in Q3 for Desktops (74.5% vs 25.5% Mercury), but the think is that AMD managed to ship more volume than the previous Quarters in a declining market.

The desktop GPU market expanded last quarter.

Which is actually good news for AMD, they got a bigger slice of a bigger pie :D
 

Pariah

Elite Member
Apr 16, 2000
7,357
20
81
It is obvious that NVIDIA did shipped more volume than AMD in Q3 for Desktops (74.5% vs 25.5% Mercury), but the think is that AMD managed to ship more volume than the previous Quarters in a declining market. That made its market share to grow vs NVIDIA. That gained didnt take place in the Steam Hardware Survey (last 3 months) for many reasons and this is why im saying that Steam Hardware Survey shouldn't be used to illustrate global desktop volume shipments.

Do you ignore anything that someone posts that doesn't fit your agenda? Samwell posted above that Mercury posts quarterly results as a 4 quarter running average. Here is the description from a Mercury Research quarterly report:

Mercury Research is
reporting the shares as a four-quarter volume-weighted average to smooth
the noise of seasonal inventory cycles and reveal ongoing share trends.

That's why the numbers differ from jpr. Not whatever you made up as the reason. The way out of line q2 results (Samwell mentioned this as well but had the wrong quarter) when Nvidia had an inventory adjustment transitioning to Pascal was factored into the q3 results. So AMD did NOT gain market share from Nvidia in q3. JPR's numbers are the more accurate results for individual quarters.

You were also wrong on the declining market part. You did acknowledge that correction by another poster since it is a positive for AMD.

Multiple posters have explained why from a mathematical standpoint, small quarterly shifts will not show up in installed userbase surveys. Again, those explanations don't promote
your agenda so you have totally ignored them and continued posting the same misinformation as if no one had already addressed your observation.
 

bystander36

Diamond Member
Apr 1, 2013
5,154
132
106
Ok my bad i didnt word it as i would like, "outsold" was not the correct word to use.

It is obvious that NVIDIA did shipped more volume than AMD in Q3 for Desktops (74.5% vs 25.5% Mercury), but the think is that AMD managed to ship more volume than the previous Quarters in a declining market. That made its market share to grow vs NVIDIA. That gained didnt take place in the Steam Hardware Survey (last 3 months) for many reasons and this is why im saying that Steam Hardware Survey shouldn't be used to illustrate global desktop volume shipments.

Understandable, it gets confusing when saying Market share, then shows sales totals. I guess I could be wrong, but that leads me to believe it's the market share of total sales.
 

AtenRa

Lifer
Feb 2, 2009
14,000
3,357
136
That's why the numbers differ from jpr. Not whatever you made up as the reason. The way out of line q2 results (Samwell mentioned this as well but had the wrong quarter) when Nvidia had an inventory adjustment transitioning to Pascal was factored into the q3 results. So AMD did NOT gain market share from Nvidia in q3. JPR's numbers are the more accurate results for individual quarters.

You were also wrong on the declining market part. You did acknowledge that correction by another poster since it is a positive for AMD.

Multiple posters have explained why from a mathematical standpoint, small quarterly shifts will not show up in installed userbase surveys. Again, those explanations don't promote
your agenda so you have totally ignored them and continued posting the same misinformation as if no one had already addressed your observation.

You seem to ignore that Im only talking about DESKTOP and you always take TOTAL GPU numbers.
I have explained several times that im only using DESKTOP results and you are still using Total Marketshare

From JPR
https://www.jonpeddie.com/publications/market_watch/

Year-to-year total GPU shipments increased 0.3%, desktop graphics decreased -4%, notebooks increased 3%.

DESKTOP Y to Y shipments DID declined (i was correct in the first place after all), AMD managed to gained share no matter if NVIDIA shipped higher Volume.

There is no agenda here, try not to get it personal and read more carefully. Also when im wrong i dont have any problem acknowledge it and re-evaluate what i said.

But in this case both JPR and Mercury shows that AMD gained MarketShare in a declining Desktop. (same as Q2)
 

antihelten

Golden Member
Feb 2, 2012
1,764
274
126
DESKTOP Y to Y shipments DID declined (i was correct in the first place after all)

No you weren't. Your previous comment was about quarter to quarter, not year to year:

but the think is that AMD managed to ship more volume than the previous Quarters in a declining market.

Also remember that the -4% number you are quoting includes both integrated GPUs and discrete GPUs. When looking at only discrete GPUs (both desktop and notebook), then the market grew by 35.6% on a quarter to quarter basis, and by 10.1% on a year over year basis.
 

AtenRa

Lifer
Feb 2, 2009
14,000
3,357
136
Also remember that the -4% number you are quoting includes both integrated GPUs and discrete GPUs. When looking at only discrete GPUs (both desktop and notebook), then the market grew by 35.6% on a quarter to quarter basis, and by 10.1% on a year over year basis.

Ahh yes you are correct, one (total Desktop + APUs) is in the start of the article and has a decline -4% Y to Y and the other (Desktop + Laptop dGPUs only) is in the end of the article and has an increase of 10.1% Y to Y.

But I was only talking about Desktop dGPUs the entire time (Mercury).