Nvidia's GPU market share hits 90% in Q4 2024 (gets closer to full monopoly)

[Zazigalka]

Like I said, I don't give a one legged, one eyed, sick rat's arse about upscaling. I don't use it, not interested. What's so difficult to understand?

why are you making it about YOU ? We're not discussing why YOU're buying one brand over the other, we're discussing why one sells better in general. you mentioned amd winning in open-source technologies, I only pointed out that fsr being open source isn't actually a tanglible advantage for current sales. if it's a win, it's a pointless one, because r7000 users don't really gain anything from it.

That's your opinion and it's ok, but it's still a device that existing and has a considerable market share.

I might have misjudged how switch sells, if you guys insist it does really well, I won't press anymore. Apparently, I'm just plain wrong.

 

[AusWolf]

why are you making it about YOU ? We're not discussing why YOU're buying one brand over the other, we're discussing why one sells better in general.

No, you're trying to prove to me that Nvidia is objectively better for some reason, even though there isn't such a thing as objectively better. Your mindset is a perfect example of why Nvidia sells more.

I might have misjudged how switch sells, if you guys insist it does really well, I won't press anymore. Apparently, I'm just plain wrong.

Personally, I wouldn't talk about the Switch on the same page together with the other consoles. Nintendo's sales largely come from their exclusives. It's not a desktop replacement the way the PS and Xbox are. It's its own universe so to speak.

 

[Zazigalka]

there isn't such a thing as objectively better.

wat ?

 

[AusWolf]

wat ?

I just said it above. There's no such thing as better in raster perf, RT, features, efficiency, OS support, price and everything else. You have to choose which of these things is important to you and base your purchase on that. If other people think other things are important, that's their choice. It doesn't mean they got an inferior product.

Edit: You might think that the Ford Fiesta is inferior to the BMW 4-series just because it's cheaper. Actually, it is smaller, easier to park, more economical and cheaper to maintain, which are advantages that outweigh those of the BMW in my eyes.

 

[Zazigalka]

I just said it above. There's no such thing as better in raster perf, RT, features, efficiency, OS support, price and everything else.

you're really claiming some ridiculous stuff. there isn't such a thing as better performance, features and efficiency ? what is shown in performance and efficiency charts then ? what is discussed in DLSS vs Xess vs FSR reviews ?

 

[AusWolf]

you're really claiming some ridiculous stuff. there isn't such a thing as better performance, features and efficiency ?

There's isn't a GPU that is better at everything all at once. I'm starting to think that you really have some reading comprehension issues.

 

[JustBenching]

There's isn't a GPU that is better at everything all at once. I'm starting to think that you really have some reading comprehension issues.

Well if someone claimed that the 4050 is better than the 7900xtx you would call him a lunatic right? Yet he can use the same argument you are using "oh it's better for me because it came in a nicer box" or whatever else subjective thing he can come up with to justify that. It's just a reality that for the 2 products you mentioned (xtx and 4080) the latter is the more complete package. You not using upscaling is irrelevant, it is an added feature (on top of FSR, since nvidia cards can also use FSR).

 

[AusWolf]

Well if someone claimed that the 4050 is better than the 7900xtx you would call him a lunatic right? Yet he can use the same argument you are using "oh it's better for me because it came in a nicer box" or whatever else subjective thing he can come up with to justify that. It's just a reality that for the 2 products you mentioned (xtx and 4080) the latter is the more complete package. You not using upscaling is irrelevant, it is an added feature (on top of FSR, since nvidia cards can also use FSR).

The XTX is equal or better in raster, it's got the 8-pin power plug, and is also cheaper. If you don't care, if you only care about RT and DLSS, even at a higher price, that's you. But to say that the 4080 is better at everything is false.

 

[JustBenching]

The XTX is equal or better in raster, it's got the 8-pin power plug, and is also cheaper. If you don't care, if you only care about RT and DLSS, even at a higher price, that's you. But to say that the 4080 is better at everything is false.

The 1650 doesn't even need a power plug, it's smaller, draws a lot less power and it's cheaper. Obviously it's better than the xtx.

 

[AusWolf]

The 1650 doesn't even need a power plug, it's smaller, draws a lot less power and it's cheaper. Obviously it's better than the xtx.

I was comparing at the same performance level. Be reasonable, man.

 

[JustBenching]

I was comparing at the same performance level. Be reasonable, man.

But you are not, the rt on the 4080 is probably on the 9900xtx level, a yet unreleased amd card...

 

[AusWolf]

But you are not, the rt on the 4080 is probably on the 9900xtx level, a yet unreleased amd card...

Yes, comparing to a theoretical GPU that you just made up is perfectly reasonable.

Here you go, by the way:

 

[igormp]

I might have misjudged how switch sells, if you guys insist it does really well, I won't press anymore. Apparently, I'm just plain wrong.

As someone has said before, it's the 3rd most sold console of all time, behind the PS2 and NDS.

 

[oxrufiioxo]

Yes, comparing to a theoretical GPU that you just made up is perfectly reasonable.

Here you go, by the way:
View attachment 378620

4080 vs 7900XTX isn't the best comparison though the 4080 super is better at basically everything right now though. Also 7900XTX is still their best selling RDNA3 card according to steam at the very least it has the most presence of any RDNA3 card and was the hardest to get at launch.





Ditching the high end around 2014 is when AMD really started to drop in market share it was respectable prior to that and has only gotten worse over time since then...



60/40 spit was super respectable

 

[AusWolf]

4080 vs 7900XTX isn't the best comparison though the 4080 super is better at basically everything right now though. Also 7900XTX is still their best selling RDNA3 card according to steam at the very least it has the most presence of any RDNA3 card and was the hardest to get at launch.



View attachment 378637View attachment 378639

Ditching the high end around 2014 is when AMD really started to drop in market share it was respectable prior to that and has only gotten worse over time since then...

View attachment 378640

60/40 spit was super respectable

That's why I keep saying that market share is an insignificant detail. You have to be profitable. There's no other goal in making business. Market share is just the favourite throwaruond word of some fans.

 

[oxrufiioxo]

That's why I keep saying that market share is an insignificant detail. You have to be profitable. There's no other goal in making business. Market share is just the favourite throwaruond word of some fans.

40% market share while being profitable would still be ideal vs breaking even with 20% or lower market share.... I've said this many times, thank god they have the Xbox/PS cuz if not it would be super dire for them at least on the gpu/gaming side.

 

[Mr_Engineer]

That alone won't be enough. Strix Halo is not monolithic, and thus won't have good power efficiency, so you won't be able to compare it to the Pro lineup.
It may be able to match the M4 Max in raw multithread performance, but will loose in anything that requires memory bandwidth, GPU and power efficiency.

I am disappointed to find out that AMD used a 5nm process for Strix Halo So for sure it will have worse efficiency than M4 series.

Regarding memory Bandwidth, the M4 Pro has 273GB/s (the M4 Max has +410 GB/s), compared to Strix Halo at 256GB/s (but includes 32MB Infinity Cache). AMD had a slide comparing rendering performance:

Not sure about GPU/gaming performance, but Strix Halo should be at least 2-3x faster than Ryzen AI 370. Which means it will also be faster than the M4 Pro in graphics.

Overall, this solution is still more efficient than having a separate CPU and dGPU laptop, and with such great performance AMD should be able to increase market share with Strix Halo against Intel / Nvidia / Apple.

 

[Bomby569]

i made up my mind i'm not buying anything under 12GB, and the 5070ti is too expensive for me to justify a purchase of a gpu. I will wait for a 5060ti with 16GB, lets see the price.

i just can't understand that 5070 with 12GB, i don't get it. Was putting more 4GB that much expensive? $10, $20 savings lol

 

[AusWolf]

i made up my mind i'm not buying anything under 12GB, and the 5070ti is too expensive for me to justify a purchase of a gpu. I will wait for a 5060ti with 16GB, lets see the price.

i just can't understand that 5070 with 12GB, i don't get it. Was putting more 4GB that much expensive? $10, $20 savings lol

Chip design, probably. For 4 GB more, you need a 256-bit bus. The 5070 only has a 192-bit one.

 

[oxrufiioxo]

Chip design, probably. For 4 GB more, you need a 256-bit bus. The 5070 only has a 192-bit one.

Thankfully GDDR7 comes in 3Gb chips so they will be able to do 18 and 24 if they choose to... With a super refresh once they are available.

 

[AusWolf]

Thankfully GDDR7 comes in 3Gb chips so they will be able to do 18 and 24 if they choose to...

Really? I didn't know, but that sounds cool!

 

[oxrufiioxo]

Really? I didn't know, but that sounds cool!

I agree especially for a 5070 super...

 

[Mr_Engineer]

VRAM Memory size seems to have stagnated for a while, a powerful card such as the RTX 5080 should have at least 24GB. Not sure why Nvidia didn't use the 3Gb chips, maybe due to low stock/availability?

Or probably they are leaving room for a cut down version of the RTX 5090, like an RTX 5080 TI with 384bit bus and 24GB memory.

 

[oxrufiioxo]

VRAM Memory size seems to have stagnated for a while, a powerful card such as the RTX 5080 should have at least 24GB. Not sure why Nvidia didn't use the 3Gb chips, maybe due to low stock/availability?

Or probably they are leaving room for a cut down version of the RTX 5090, like an RTX 5080 TI with 384bit bus and 24GB memory.

I don't think they are available yet....

 

[Bomby569]

I don't think they are available yet....

i google it, at least samsung is only starting production this year

Samsung introduces GDDR7 memory in 3GB modules — one and a half times larger and twice as fast

Samsung has announced the industry's «first» 24-gigabit (3 gigabytes) GDDR7 memory for graphics cards. It supports speeds of over 40 Gbps.

itc.ua