These are Q3 daily numbers for three days: 31%, 24%, 27% Q1 = 18.1% Q2 = 20.6% So Q3 off to a bang for AMD, with OEMs building Ryzen systems now, not just DIY. So far looking like high 20s for Q3, doubling the gains from Q1 to Q2.

Where are the figures from?

This article is deceptive at best. How could AMD capture 10% when they weren't even selling a chip under $329 until half way through the quarter? The problem with PC mark is its a self selected survey of not just enthusiasts, but enthusiasts who specifically benchmark their PC's using PC mark software. That is not a good indicator of the overall market. The true market share gain is likely somewhere between 1-2% https://seekingalpha.com/article/4085523-amd-gain-10-percent-market-share-quarter

Well, since people keep saying that "this means that the market share of passmark users increased for AMD", I'll try to correct this absurd excuse. Market research are based on a small amount of people, for instance, for a country with 18 million citizens, it will take about 1500 people to know how your product will perform among the 18 million, the margin of error is VERY small. So, AMD did take a lot ( 2-5% is a lot of market share ) of market share from Intel, even if the report comes from a benchmark, steam or pornhub. Regards

Perhaps you can state how reliable an indicator this benchmark presents. There must be a reason you put a foundation in it. As you are quoting %, how reliable in % is it? regards

Are there credible articles about that?

Being that gamers using Steam make up a small percentage of the PC market, it should be obvious to anyone with half a brain that Steam hardware survey does not produce accurate numbers.

I meant the gap of market share between Intel and AMD, which according to that graph will get smaller as AMD gains and Intel loses.

That graph is based on PassMark benchmark result submissions. Being that the figures are based on benchmark result submissions, the numbers only mean that more AMD users are submitting results now, compared to last quarter. The figures are completely meaningless in regards to actual market share or sales.

I mean, is not that crazy to take this information as reliable, it's not extremely accurate, but it is reliable. AMD did take an important share from Intel, it's not 10%, but they did take around 2-5%

Can you prove all of what you just stated as fact?

I mean, is not that crazy to take this information as reliable, it's not extremely accurate, but it is reliable. AMD did take an important share from Intel, it's not 10%, but they did take around 2-5%

These results mean nothing in regards to actual market share. It's based solely on people submitting benchmark results. All it shows is that more people are submitting benchmark results for AMD hardware now, than they were last quarter....or fewer Intel users are submitting results. Considering how long the current Intel hardware has been available, it's not inconceivable to believe that submissions from Intel users has simply dropped off.

Last week I finally got steam HW survey after so many years of ignoring me. My Fury X and 2 new SSDs are finally accounted for. 😀 Now I can get something new, to have it ignored for another 4~5 years.

I haven't done one in few years. Last time I did it I had 0 ssd's, 290x, 3770k and something like that.

I don't think I got a steam hardware survey request since 2010 or so.

From what I've heard manufacturing cost is higher on TR boards.

You're most likely correct, those 64-PCIe lanes and the 4094-pin slot won't be cheap. On the other hand, AMD is talking constantly about a lower total cost of ownership with EPYC and TR, and X299 seems to be a difficult platform to build for as a motherboard manufacturer, just by the sheer amount of features you have to have toggles for (it goes from 16x PCIe lanes up to 44, and from dual channel memory to quad, and then you have the RAID silliness on top). TR boards might be more expensive material wise, they seem more straightforward construction/maintenance wise. We'll see. I too believe they'll be a bit more expensive than the X299 boards, although some might not even have a chipset.