I’m not planning on buying a GPU now, but nonetheless I never really understood how GPU’s are being compared to each other.
I have a specific problem that I have a “custom” VTX3D GPU, so it is impossible to find it on “compare GPU” websites, also vtx3d doesn’t have a website anymore, so I have to learn how they work in order to not get something less powerful in the future if I decide to upgrade.
Its a Radeon HD 7730 2gb DDR3. The vendor probably just bought the same board all the other vendors bought… they just ask for the board to made with this look of board, or that heat sink … just to get it unique…
VTX3d don’t design the electronics working inside of it, so its limitted to what AMD ATI Radeon has designed.
AMD provide the driver package with the ability to put the brand name on the software too…
You can check that all Radeon HD 7730 2gb DDR3 are similar in performance , right ?
Especially if you stick to a desktops and not any sort of special version of it ,if any.
Hm…but what is the difference in performance for a 1gb gddr5 and 2gb gddr3 version? The one I have is the 2gb gddr3 and most comparison websites just list the 1gb gddr5 version, so I can’t really figure out what I am even suposed to look at.
With Cpu’s it’s generally easy, you mostly look at cores, threads and the nm technology number (14, 22, 32,etc.) and that solves 99% of your questions about the cpu, but here there’s a texture fill rate, pixel fill rate, memory mhz, cuda/stream procesors, clock mhz, just a gazillion of different things and I have no idea which one is the most important.
These things are part of a larger system that works in concert to do its job. Different tasks will stress different parts more than others.
For instance, Nvidia graphics cards generally do better than AMD cards when running modern games. However, AMD cards work better if you want to mine cryptocurrencies like Etherium.
Things like higher bandwidth or faster memory may count for a lot or not at all.
All that said a modern, bleeding edge GPU will almost certainly be faster than your old GPU.
If you are interested in a new card I would benchmark your current card then buy a new card and run the same benchmarks. If the difference is not enough return the new card for a refund.
Looking at the GPU Hierarchychart for Tom’s Hardware, the 7730 DDR3 is 17 tiers in performance down from the top tier.
For reference, a card like the RX 460 (~$110) is about 7 tiers from the top.
I’d doublecheck that your motherboard can support anything new. I think PCI-e 3.0 is backwards compatible with the PCI-e 1.0 socket(which I believe the 7730 has), but you’ll want to verify.
The extra gig won’t help you one bit unless you fill it. Your software has to make use of that extra gig… basically modern games, or 4K video in a program like Premiere Pro.
Also, it won’t help a poor performing card, as the card won’t have enough performance to process video data to make use of more than the lesser gigabytes. Thats why memory size is increased with GPU performance, and there are two tiers. The ordinary folks card, and a bit more expensive one with double the Gig’s , and thats the maximum size GPU performance could make use of.
Even the new budget GT 1030 (which is awesomely low powered) gets better scores than your card. Since your GPU officially maxes out at 1GB, the additional memory is almost certainly unhelpful.
You can always do your own benchmarks with stuff like PassMark or similar, and then use that to compare to the benchmarks of the other cards they list. The GPU tests are designed to minimize influence from other components.