Qcumber Posted March 12 Posted March 12 (edited) Just got my new 5070ti (MSI Ventus 3x OC) so I thought a new thread might be useful. I've upgraded from a 4070 so was expecting a reasonable improvement. It's early days but the jump is much better than I expected. Synthetic benchmarks (Steel Nomad and Furmark) show about a 60% bump over the 4070 which is consistent with reviews at 4k. It's always tricky evaluating performance when fps is maxed out so I tend to look at %GPU use. I use the Syria F-16 free flight as one of my performance checks. With my previous setup (5800x3d and 4070) GPU was maxed out when flying over Tripoli. With the same DCS settings my new setup (9800x3d and 5070ti) the GPU load maxed out at about 65%. (Does anyone have ideas how best to give a comparative DCS benchmark?) I have managed to overclock it to +600 core and +2000 in MSI afterburner. This is stable in Furmark and after a couple of hours of intensive missions in DCS. In Furmark this gives me +17% over base performance. I realise this is synthetic and not necessarily representative but it's still a very nice boost. The other thing I noticed is that DLSS 4 appears to run better. It could be that I am able to push resolution up and can maintain better latencies. It it could be that 50xx GPUs are better optimised!? Edit: I have since discovered that the +600 core boost is not stable and causes core throttling. +400 is stable in DCS with DLSS4 and gives me +8% in Steel Nomad. Edited March 14 by Qcumber 9800x3d; rtx5080 FE; 64Gb RAM 6000MHz; 2Tb NVME; Quest Pro (previous rift s and Pico 4).
The_Nephilim Posted March 14 Posted March 14 On 3/12/2025 at 3:27 PM, Qcumber said: Just got my new 5070ti (MSI Ventus 3x OC) so I thought a new thread might be useful. I've upgraded from a 4070 so was expecting a reasonable improvement. It's early days but the jump is much better than I expected. Synthetic benchmarks (Steel Nomad and Furmark) show about a 60% bump over the 4070 which is consistent with reviews at 4k. It's always tricky evaluating performance when fps is maxed out so I tend to look at %GPU use. I use the Syria F-16 free flight as one of my performance checks. With my previous setup (5800x3d and 4070) GPU was maxed out when flying over Tripoli. With the same DCS settings my new setup (9800x3d and 5070ti) the GPU load maxed out at about 65%. (Does anyone have ideas how best to give a comparative DCS benchmark?) I have managed to overclock it to +600 core and +2000 in MSI afterburner. This is stable in Furmark and after a couple of hours of intensive missions in DCS. In Furmark this gives me +17% over base performance. I realise this is synthetic and not necessarily representative but it's still a very nice boost. The other thing I noticed is that DLSS 4 appears to run better. It could be that I am able to push resolution up and can maintain better latencies. It it could be that 50xx GPUs are better optimised!? can you do a firestrike benchmark and post your score? 1 Intel Ultra 265K 5.5GHZ / Gigabyte Z890 Aorus Elite / MSI 4070Ti Ventus 12GB / SoundBlaster Z SoundCard / Corsair Vengance 64GB Ram / HP Reverb G2 / Samsung 980 Pro 2TB Games / Crucial 512GB M.2 Win 11 Pro 21H2 / ButtKicker Gamer / CoolerMaster TD500 Mesh V2 PC Case
Qcumber Posted March 14 Author Posted March 14 17 minutes ago, The_Nephilim said: can you do a firestrike benchmark and post your score? Will do. I have run Steel Nomad as I figured this was newer and therefore better but will test Firestrike too. 9800x3d; rtx5080 FE; 64Gb RAM 6000MHz; 2Tb NVME; Quest Pro (previous rift s and Pico 4).
The_Nephilim Posted March 14 Posted March 14 1 hour ago, Qcumber said: Will do. I have run Steel Nomad as I figured this was newer and therefore better but will test Firestrike too. Well Nomad is DX12 and Firestrike is a DX11 test which is more based on DCS.. I would be interested to see what you score vs my system.. 1 Intel Ultra 265K 5.5GHZ / Gigabyte Z890 Aorus Elite / MSI 4070Ti Ventus 12GB / SoundBlaster Z SoundCard / Corsair Vengance 64GB Ram / HP Reverb G2 / Samsung 980 Pro 2TB Games / Crucial 512GB M.2 Win 11 Pro 21H2 / ButtKicker Gamer / CoolerMaster TD500 Mesh V2 PC Case
Qcumber Posted March 14 Author Posted March 14 Rtx 5070ti VR performance 9800x3d; rtx5080 FE; 64Gb RAM 6000MHz; 2Tb NVME; Quest Pro (previous rift s and Pico 4).
Qcumber Posted March 14 Author Posted March 14 4 hours ago, The_Nephilim said: can you do a firestrike benchmark and post your score? Here are my Firestrike results. Single test at each configuration. I plan to do repeat tests and average them. Core Mem Overall GPU Physics Combined Base 0 0 51588 67955 37721 22879 OC 400 2000 53218 72775 36590 22804 % n/a n/a 103% 107% 97% 100% 9800x3d; rtx5080 FE; 64Gb RAM 6000MHz; 2Tb NVME; Quest Pro (previous rift s and Pico 4).
The_Nephilim Posted March 15 Posted March 15 15 hours ago, Qcumber said: Here are my Firestrike results. Single test at each configuration. I plan to do repeat tests and average them. Core Mem Overall GPU Physics Combined Base 0 0 51588 67955 37721 22879 OC 400 2000 53218 72775 36590 22804 % n/a n/a 103% 107% 97% 100% 1 Intel Ultra 265K 5.5GHZ / Gigabyte Z890 Aorus Elite / MSI 4070Ti Ventus 12GB / SoundBlaster Z SoundCard / Corsair Vengance 64GB Ram / HP Reverb G2 / Samsung 980 Pro 2TB Games / Crucial 512GB M.2 Win 11 Pro 21H2 / ButtKicker Gamer / CoolerMaster TD500 Mesh V2 PC Case
BeefSupreme Posted Sunday at 07:57 AM Posted Sunday at 07:57 AM I just picked up this card, will have it in a few days. What VR setup are you rocking?
Qcumber Posted Sunday at 08:01 AM Author Posted Sunday at 08:01 AM Just now, BeefSupreme said: I just picked up this card, will have it in a few days. What VR setup are you rocking? I ended up getting a 5080 and sending the 5070ti back. It's a good card but I managed to get a 5080 FE for oy £150 more than the 5070ti. The 5070ti works very well in VR. I use a Quest Pro with eye tracking. It can manage 72 fps in most situations. 9800x3d; rtx5080 FE; 64Gb RAM 6000MHz; 2Tb NVME; Quest Pro (previous rift s and Pico 4).
Recommended Posts