Well with generally no feedback and very little info on 5050 GPU's I snagged a 5060Ti 8GB for $30 under MSRP with an additional $30 rebate. More than I planned to spend but seriously better than a 5050 on paper. Question is will the PCIE5 x8 interface hamper it for FAH on a PCIE3 mobo? I can swap a few things around and get it on a PCIE4 mobo but I'd rather not. Things I've read in the past suggest the PCIE lanes don't matter so much for FAH but with the reduction in bandwidth from PCIE5 vs 4 vs 3 do you think it'll be a big hit?
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
-
Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.
NVM RTX 5050
- Thread starter Toconator
- Start date
- Joined
- Oct 7, 2012
- Messages
- 425
I have a 5080 on an Ivy Bridge Mobo. No noticeable impact for FAH of the pcie bandwidth that is sufficient. But there is a variable impact of the CPU depending on the projects themselves, notably on how the researcher set up checkpoints and sanity checks.
Overall 10-20% PPD less between an Ivy bridge quand core and a 5600X, in my experience.
Worst on certain projects with large number of atoms notably solvent. For example for P18262 currently under beta test:
- 13,8 M PPD on 5080 + Ivy bridge + HDD, checkpoints take around 45 seconds
- 24 M PPD on 5080 + 9950X + SSD 2671 Mb/s, CPU folding 30 cores in parallel, checkpoints take around 2 seconds
- for reference 38,3 M PPD on 4090 + 5950X + SSD 6200 Mb/s, no CPU folding in parallel, checkpoints take 7 seconds
This illustrates the importance of the CPU on certain large projects, and its superior importance vs SSD speed.
Overall 10-20% PPD less between an Ivy bridge quand core and a 5600X, in my experience.
Worst on certain projects with large number of atoms notably solvent. For example for P18262 currently under beta test:
- 13,8 M PPD on 5080 + Ivy bridge + HDD, checkpoints take around 45 seconds
- 24 M PPD on 5080 + 9950X + SSD 2671 Mb/s, CPU folding 30 cores in parallel, checkpoints take around 2 seconds
- for reference 38,3 M PPD on 4090 + 5950X + SSD 6200 Mb/s, no CPU folding in parallel, checkpoints take 7 seconds
This illustrates the importance of the CPU on certain large projects, and its superior importance vs SSD speed.
- Joined
- May 7, 2007
- Messages
- 13,419
Well stated, I have seen very similar results as well going from Ivy Bridge and other CPUs from 2012, to modern CPUs from 2018-present.
The GPUs are very capable, but those checkpoints on older processes, you are right, they take so much longer and make a large impact on the time of processing the tasks.
Good info N_o and perhaps very useful in future as GPU's trickle down thru my boxen as I have an Ivy Bridge 3770K not in my sig anymore because it exceeds the line limit for sigs. Was thinking of the 8600K Server boxen as I have a heat issue with 2 GPUs in it but now I see the TDP of 5060Ti is 180W so it's not going to fix my problem. My 3060 is a 3 fan unit so I'll put it in the Server and see if it won't throttle with 2 GPUs. Might take a bit more GPU juggling to max out my ppd as I notice the GTX1660 Super sometimes exceeds the RTX2060 which is hitting it's temperature threshold most of the time, especially with core 0x22 WU's. Not all of these boxen are going to be running 24/7 like right now tho, I'm just seeing how to best set-up during challenges. Don't mind the extra warmth in winter but the power bill$ are killer.