AMD Big Navi Performance Claims Compared to TPU's Own Benchmark Numbers of Comparable GPUs

[medi01]

He knew Steam already fixed the counting algorithm and he used old data trying to discredit Steam Survey.
What else is new ?

He mentioned Steam fixed some but not all issues reported by AMD.
He wasn't specific about what issues were not addressed.

Steam figures contrast with other sources, major retailer stats being one of them, I don't get why you don't let that dead horse just be.
Mindfactory had AMD (unit) market share between 29%-39% for this year, with 5700 series and 2070 series being the most popular models.

 

[nguyen]

He mentioned Steam fixed some but not all issues reported by AMD.
He wasn't specific about what issues were not addressed.

Steam figures contrast with other sources, major retailer stats being one of them, I don't get why you don't let that dead horse just be.
Mindfactory had AMD (unit) market share between 29%-39% for this year, with 5700 series and 2070 series being the most popular models.

Let me tell you that you don't understand the fundamental difference between Steam data and Mindfactory data.
Steam data represent what GPU are being used by Steam users.
Mindfactory data are number of GPU being sold.
Steam is the absolute percentage while Mindfactory is the relative change, get it ?

Let say there are 100 millions dGPU circulating the market with Nvidia 80% and AMD 20%, an additional 1 million GPU being introduced into the market every month with Nvidia and AMD selling equal amount of cards would make that market share barely change (79.7% vs 20.3%), get it ?

So stop comparing apple to orange.

Also Mindfactory is a single retailer, that is like asking who would win the POTUS election by counting votes in a single city.

 

[medi01]

Steam is the absolute percentage while Mindfactory is the relative change, get it ?

Shocking story indeed, except Steam shows changes too.

 

[Adam Krazispeed]

I could live with that if the price is belowe the 3080

sorry techpowerup but this is my bet...... big navi XT/XTX 6900 XTX/XT that is 2400Mhz game clock possibly 2.6/2.7ghz boost (MAX) boost clocks.. 135-137% relative Performance......sorry i got more info i cant share yet.... hehe this is all i can do.. for now!!!

 

[hat]

2.4GHz?

 

[medi01]

2.4GHz?

Overhyping is a method to make launch of an excellent product a disappointment.
One of the FUD arsenal moves.

36CU RDNA2 chip (PS5) base is 1.8GHz and it boosts to only 2.23GHz, there is no way 80CU chip would have "base clock" of 2.4.

56CU RDNA chip (XSeX) clocks at 1.825Ghz (no boosting).

2.4+ GHz as base is just FUD campaign.

 

[Zach_01]

Overhyping is a method to make launch of an excellent product a disappointment.

I agree! There is no need for this...

36CU RDNA2 chip (PS5) base is 1.8GHz and it boosts to only 2.23GHz, there is no way 80CU chip would have "base clock" of 2.4.

56CU RDNA chip (XSeX) clocks at 1.825Ghz (no boosting).

2.4+ GHz as base is just FUD campaign.

You actually cannot base the clock assumptions upon consoles. High-end dGPUs have a power budget higher than an entire console that includs a CPU and other stuff too.
I'm not saying that an 80CU Navi21 will boost to 2.7GHz. Maybe boosting to 2.3~2.4GHz is viable.

 

[Adam Krazispeed]

ssed on Techpowerups specs and bench

 

[INSTG8R]

ssed on Techpowerups specs and benchView attachment 172230

Problem with that is all those cards were using completely different settings, API and benchmark scenarios. All they have in common is the same game and same resolution

 

[Adam Krazispeed]

i modified this to where i belive the 6900xt/xtx will sit in the stack on release

 

[INSTG8R]

i modified this to where i belive the 6900xt/xtx will sit in the stack on release

I like your optimism. I’m total Red Team but I have to remain sceptical. The leaked numbers that are floating around look promising but until they turn into benchable performance numbers they don’t mean much.

 

[TheoneandonlyMrK]

Let me tell you that you don't understand the fundamental difference between Steam data and Mindfactory data.
Steam data represent what GPU are being used by Steam users.
Mindfactory data are number of GPU being sold.
Steam is the absolute percentage while Mindfactory is the relative change, get it ?

Let say there are 100 millions dGPU circulating the market with Nvidia 80% and AMD 20%, an additional 1 million GPU being introduced into the market every month with Nvidia and AMD selling equal amount of cards would make that market share barely change (79.7% vs 20.3%), get it ?

So stop comparing apple to orange.

Also Mindfactory is a single retailer, that is like asking who would win the POTUS election by counting votes in a single city.

Pot, kettle, black, you see steam stats as that golden single source your berating others for seeing In mindfactory.

Neither can be the last word clearly.

 

[medi01]

I agree! There is no need for this...


You actually cannot base the clock assumptions upon consoles. High-end dGPUs have a power budget higher than an entire console that includs a CPU and other stuff too.
I'm not saying that an 80CU Navi21 will boost to 2.7GHz. Maybe boosting to 2.3~2.4GHz is viable.

If 80CU could have 2.4Ghz base, then 56CU will definitely go beyond 1.8Ghz.