Hi! I know this is old and you probably already got a card, BUT, just for the sake of answering and not leaving others in the abyss that may arrive at this post, I want to reply.
Disclaimer- I am in no way an expert and also am just like you as far as interested in these features.
So, in my journey of looking for an upgrade myself, I gained a lot of good insight that I think will help you, or anyone at this level, and would like to share it with the community.
For example, I saw a Redditer answer the PCIe 2.0 v.s. PCIe 3.0 question like this:
*note that while this is a reply about Gaming, the theoretical resemblance applies.
The theoretical speeds of the PCIe specs matter very little when it comes to gaming, and even less for mining.
The number of lanes are either 2x, 4x, 8x or 16x (this is seperate to PCIe 3.0 / 2.0). You can tell the number of lanes by the length of the slot (tiny slot = 2x, full size slot = 16x).
PCIe 2.0 at 16x is 95%+ performance level as PCIe 3.0 at 16x (usually 99%+)
PCIe 2.0 at 8x should still give 90%+ performance.
PCIe 2.0 at 4x however gives around 85-90% performance depending on the game, so you might notice a little drop in some games but nothing major. Still very much useable.
Source: have gamed on PCIE x4 connection and a gtx 1060 with no noticeable downsides. Comparison Chart for multiple games: [https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_1080_PCI_Express_Scaling/5.htm
Also, I learned about the compatibility between the 2 PCI’s hardware standard, in the same thread, that said:
PCIe 3.0 runs at nearly double the speed, but is still totally forward/backward compatible so 2.0 boards will work with 3.0 cards and vice versa, and the real world performance difference is pretty tiny.
Other comments read:
3.0 over 2.0 is a ~2fps difference, so no big deal
- - And that’s with the fastest GPU’s out there. The 1050ti is nowhere near powerful enough to max that PCIe 2.0 link out.
Well the 3.0 slot has more bandwidth, but that doesn’t have an effect on performance. I have a 1060 in a 2.0 slot and it runs just fine.
I posted the tidbits here, in case this should suffice for the answers anyone needs. However, you can check out the entire thread if you like, at the following link: https://www.reddit.com/r/buildapc/comments/6of16d/pcie_20_and_pci_30_difference/
I hope this helps anyone, and I’m certain that all this knowledge has helped me reach a conclusion on what GPU cards to look into. For me, it was great info because of a budget I must adhere to, and this helps me determine if buying something slightly older, or cheaper, would ‘make or break’ my needs.
For example, I found an Nvidia card with 1300+/- Cuda cores, and higher clock speed with 4gb, for PCI 2.0, for 2 times less the price of another card for PCI 3.0 with only 800+/- Cuda cores and also 4gb. So, huge performance difference between cards, with not too bad performance difference between PCI standards, I’m going with the decision based on the cards and the price.
This is not to say that PCI 3.0 does not have significant advantages over 2.0. We are just comparing for the sake of beginner to intermediate level, quality performance.
The obvious way to go, if you are already established and well-advanced, and at a level that budget is nottoo much of a concern, then by all means go ‘balls to wall’!
My only caveat is…
Just make sure to look at the connections they come with, and ports, also their size, to compare to the form factor case limits of the length, and that everything fits- including with your PSU, CPU cooler heights, etc.
Remember…
Doc Brown’s invention wasn’t marvelous and worked because of how fast it got to 88mph. It worked WHEN it hit 88mph.
Haha that’s my little quote to ponder about.
Always compare your final picks in the end directly on Nvidia sites, or other reputable Tech specs sites. And the best source is always the datasheet provided on Nvidia sites.
Have a nice day, to all!
