Why did ATI fail to keep up with NVIDIA? What factors contributed to their inability to outperform the green competitor, ultimately leading to their acquisition by AMD?

>>11838060 (OP)
Cuz gay-ti a shit

Red is a hot color, which is unappealing to g*mers who like being cool and comfy, especially during summer which is when they do a lot of their g#ming and upgrades

>>11838060 (OP)
>Why did ATI fail to keep up with NVIDIA?
Confusing naming schemes no one could keep track of, awful barely-functional drivers, poor engineering that led to lots of overheating, low compatibility, fewer bells and whistles like Nvidia's fancy AA/filtering/particle/physics shit. ATI was just always low quality, plain and simple. Barebones budget cards people only put up with because they couldn't afford Nvidia.

Wasn't ATI once caught cheating on benchmarks? Where if their drivers checked if the filename they were running was called "quake3.exe", commonly used for benchmarks at the time, they would override the mipmap settings with lower quality options so they would score higher? I recall someone noticed that just simply renaming the exe file something else exposed what they were doing.

Or was that after AMD bought them?

>>11838191
That's just how graphics drivers work because video game developers are awful people.
Games based on the various Quake games are a particularly common example. Having to rename sin.exe to quake2.exe to get the game running on newer systems because the graphics driver no longer has the game-specific fixes for that particular game as the list gets cut down over time.
Carmack has talked about this a few times, how even nearly 30 years later modern GPUs still have to implement special fixes to get GLQuake to ever start up at all.

>>11838201
It was not "just how graphics drivers work", it was a huge controversy at the time. They were basically cheating in benchmarks, intentionally lowering the image quality if they saw you running Quake to make their cards perform better in benchmarks.

https://web.archive.org/web/20231129145604/https://techreport.com/review/how-atis-drivers-optimize-quake-iii/

Yes it's common to have game-specific fixes in drivers now, but not so back then, and this wasn't a "fix" but intentionally overriding image quality settings to lower ones on a game that everyone was using as the main benchmarking tool at the time specifically to cheat in benchmarks, it was not any kind of optimization or fix.

>>11838118
I’m a zoomer, so correct me if I’m wrong but wasn’t the driver issue mostly a problem starting with GCN up to RDNA 1?

From what I’ve read in archived articles about PC hardware and gaming from that era, ATI’s drivers weren’t actually that bad. In fact, the features they offered, like image quality and anti-aliasing, were often better than what NVIDIA had at the time.

Partially unrelated, but I used to have a birman cat who looked like that, I miss her.

>>11838353
:' (
<3

>>11838250
I'm not well versed in the issue but I do know they had driver issues dating all the way back to Mech Warrior 2

Back in the day cheap GeForce were just generally better.
You would have to pay extra for a mid performance.

>>11838060 (OP)
Their market was exactly the same as AMD's target audience: minimum barely functional viable product for the low end that cannot afford the premium. They probably should have full time team of engineers to design and develop their products instead of being the part-time side-gig option for the one solo engineer primarily working at their competitors. I remember in performance metrics it was nVidia, but for static image quality ATi had it, but eventually nVidia took that crown too. Even their best cards has this issue with the glue/epoxy expanding removing contact with heatsinks, requiring thicker and thicker thermal pads to transfer heat from the chip to another surface. I was on team red not by choice, but circumstance.

>>11838212
I remember this. I knew this was true, but it seems like all this information was scrubbed from the internet. I even remember this happening more than with just Quake, but also with MadOnion's tech demos. AMD/ATi probably spent more on publicity damage control than actually making a good product.

>>11838060 (OP)
There was a time when it looked like they were getting their shit together, starting with the Radeon 8500, though that was kinda held back by shit drivers (which were apparently fixed later). The 9000 series were a great success, and it's the one time they utterly and completely BTFO'd Nvidia, who were caught stumbling with the FX series. That series proved to be lightning in a bottle, though, and once Nvidia got it together for the next series, ATI just wasn't able to get the upperhand again.