What's the point of spending more than $500 on a GPU if everything you'll do on it is just gaming?
I know how redditor gamers and indian ITs love to praise latest amd series for their price to performance, vram and everything, but would it be worth it if it's fucking useless for anything besides gaming compared to a nvidia gpu? because AMD had decades in this business yet they still couldn't come with something that remotely competes with CUDA, stingy motherfuckers don't even want to spend money or invest on that direction, it's like they see nvidia having a monopoly and they're just ok with it.
Don't be fucking retard and mention rocm, rocm is fucking garbage both in support and availability it's really more like a cuda from temu, migrating from windows to linux and only sticking with 5000/6000 gpus just for propietary acceleration that isn't even as good as cuda, you're fucking yourself if you do, the software and hardware that have rocm and actually use rocm can be counted on one hand.
In this day and age you won't be able to find any gpu accelerated task that isn't limited to cuda or won't benefit from cuda, and most of the time it's the first case (either cuda or cpu accel)
Media workloads, ML workloads, gay-i, all of them highly benefit from cuda, hell at one point I thought that python was fucking owned by nvidia due to how much cuda you see in python.
want a useful script/tool that does things like media upscaling or resampling audio? cuda
want to use cool features for your video/photo editing software like davinci or adobe slop? cuda
got CADs or modeling/EDAs or other engineering related shit with gpu workloads? cuda
want to get in zoomer Gay-i and ML compute with python? cuda
if you don't have cuda you're literally fucked, and like I said most of the times I encountered cuda support the alternative wasn't even non propietary regular gpu accel, it was either cuda or cpu acceleration, the task that it would take a gpu to complete in a minute becomes a behemoth that lasts for hours with hilarious RAM requirements.
TL;DR nvidia will always be able to manipulate this market and can keep pushing new series, price their cards however they want and they would still sell like hot cakes, people think it's just corruption but there's a good reason for that, they have a monopoly over software support which they worked and dumped money on all these years while amd was always busy playing catchup in hardware.
I know how redditor gamers and indian ITs love to praise latest amd series for their price to performance, vram and everything, but would it be worth it if it's fucking useless for anything besides gaming compared to a nvidia gpu? because AMD had decades in this business yet they still couldn't come with something that remotely competes with CUDA, stingy motherfuckers don't even want to spend money or invest on that direction, it's like they see nvidia having a monopoly and they're just ok with it.
Don't be fucking retard and mention rocm, rocm is fucking garbage both in support and availability it's really more like a cuda from temu, migrating from windows to linux and only sticking with 5000/6000 gpus just for propietary acceleration that isn't even as good as cuda, you're fucking yourself if you do, the software and hardware that have rocm and actually use rocm can be counted on one hand.
In this day and age you won't be able to find any gpu accelerated task that isn't limited to cuda or won't benefit from cuda, and most of the time it's the first case (either cuda or cpu accel)
Media workloads, ML workloads, gay-i, all of them highly benefit from cuda, hell at one point I thought that python was fucking owned by nvidia due to how much cuda you see in python.
want a useful script/tool that does things like media upscaling or resampling audio? cuda
want to use cool features for your video/photo editing software like davinci or adobe slop? cuda
got CADs or modeling/EDAs or other engineering related shit with gpu workloads? cuda
want to get in zoomer Gay-i and ML compute with python? cuda
if you don't have cuda you're literally fucked, and like I said most of the times I encountered cuda support the alternative wasn't even non propietary regular gpu accel, it was either cuda or cpu acceleration, the task that it would take a gpu to complete in a minute becomes a behemoth that lasts for hours with hilarious RAM requirements.
TL;DR nvidia will always be able to manipulate this market and can keep pushing new series, price their cards however they want and they would still sell like hot cakes, people think it's just corruption but there's a good reason for that, they have a monopoly over software support which they worked and dumped money on all these years while amd was always busy playing catchup in hardware.
Last edited:
