Hello. I’m hoping the hive mind here can give me some advice. I have a half decent video card (gtx 1050) running ~35-45 FPS at 1080p in ultra on an older system (core 2 quad / 4gb ram). I noticed that if I drop to 720p or I raise to 1440p it doesn’t seem to make any difference at all on the frame rate. The only thing that seems to affect the frame rate is how many other people are on Zwift at the moment (busy days definitely seem to correlate with a lower frame rate) Does this mean I’m somehow limited by the rest of the system and my video card isn’t the bottleneck? Thanks!
Which 1050 do you have, The GTX1050 2GB or the GTX1050ti 4GB? Your GPU is probably the most important hardware in your PC as far as Zwift is concerned. From the info you posted it seems It is your GPU is the bottleneck.
Check your vertical sync settings in the nVidia control panel. If you have it on then your maximum frame rate is the refresh rate of your monitor, often 60 fps.
I agree, resolution on a Windows PC is not the determining factor for your FPS.
Zwift will implement a 4K shadow res at any resolution @ Ultra setting and that’s the first bottleneck. The second and more important bottleneck is what you described, traffic. From what I can see there is a direct relation in your graphics card ability to draw the next frame, as there is in Zwift’s servers ability to give you that frame.
The best case scenario gives you a GPU that is strong enough to draw any frame but also fast enough to accept that frame as late as possible (in traffic) and still draw it. The faster your GPU, the better you will be able to deal with latency from the Zwift servers.
As far as a GTX 1050 goes, good luck at any Ultra setting, just not strong enough to push Zwift at max. At a lower setting with a lower shader res, the 1050 will put out over 60 fps.
Thanks everyone! I’m using the GTX 1050 2gb (non ‘ti’ version, but it is the’OC’ version) with v-sync off. I’ve tried pushing down the shadow res and doing other things to force it down to ‘high’ or ‘medium’ settings and the frame rate doesn’t improve. I’m still curious about whether the overall system itself, which is old and still uses pci express v1.0 is somehow holding it back. But based on what everyone is saying, it sounds like the consensus is on the video card being the limiting factor, right?
version Introduced Line
×1 ×2 ×4 ×8 ×16
1.0 2003 8b/10b 2.5 GT/s 250 MB/s 0.50 GB/s 1.0 GB/s 2.0 GB/s 4.0 GB/s
2.0 2007 8b/10b 5.0 GT/s 500 MB/s 1.0 GB/s 2.0 GB/s 4.0 GB/s 8.0 GB/s
3.0 2010 128b/130b 8.0 GT/s 984.6 MB/s 1.97 GB/s 3.94 GB/s 7.88 GB/s 15.8 GB/s
4.0 2017 128b/130b 16.0 GT/s 1969 MB/s 3.94 GB/s 7.88 GB/s 15.75 GB/s 31.5 GB/s
5.0 expected in
Q1 2019  128b/130b 32.0 GT/s[ii] 3938 MB/s 7.88 GB/s 15.75 GB/s 31.51 GB/s 63.0 GB/s
Yes, if your motherboard only supports v1.0 you are already limited beyond your GPU capabilities. You’re in 2003-2006 just trying to hang on.
Have you looked at the Zwiftalyzer benchmarks to see what other people’s Zwift rigs are doing?
The clock speed of your old core 2 quad CPU is likely a bottle neck to your GPU not running at its full potential.
Hey Mike. You are totally right. I moved the card into an i5 system and it almost doubled my frame rate! I’m getting anywhere from 50-80 using 1080 ultra depending on the world and how busy it is. I still find that I change from 1080 to 1440 without any serious hit to my frame rate so I’m assuming I’m still limited somewhere other than the video card, but I’m going to leave it be as I’m happy with the what I’m seeing. At this point the money would be better spent elsewhere. Thanks everyone!
I had the same experience with a Core 2 Duo E8400 Wolfdale 3.0 Ghz. The GPUs (GTX 750 Ti, GTX 960, GTX 970 and AMD 7870) were all underutilized with one of the CPU cores maxed out at 100%. The 970 was just getting tickled at ~30% utilization. I get the same result with a quad Xeon X5672 Westmere EP at 3.2 GHz. Those CPUs are just too old now to push modern GPUs, but do make great beater machines for the shed. When I moved up to a Pentium G3258, and Core i5 4690K, at stock clock speeds, I was able to run the GPUs closer to 100% utilization. I actually keep v-sync on now to cap the FPS at 60 @ 1080 ultra with the 970 to deliberately cap the GPU utilization to keep it cool and prolong its life. (I’m not made of money). I used CPUID HW monitor to see the usages and temperatures. Generation and clock speed of the CPU does makes a difference with discrete GPUs. I’d generally recommend 4th generation Intel CPU (haswell refresh) or newer, with higher clock speed over more cores, with a minimum of 2 cores. The sweetspot of performance to cost right now (Feb 2019) is the Core i5-7500 Kaby Lake Quad-Core 3.4 Ghz and the GTX 1050 Ti.
Thanks again Mike! I forget which of the i5’s I have, but I know it has four cores because I looked it up when I bought it. I also tried re-installing Win 10 from scratch and notice that seems to have boosted my frame rates. The i5 that I bought was a used machine and had been upgraded from Win 7, then migrated from HDD to SSD and that ‘fresh’ install seems to have helped. In fact now I’m getting 60fps without a sweat. Maybe it was just the installation of all new drivers and reformatting of the hard drive. I also set v-sync to on as I don’t see the benefit of the gpu pushing the frame rate any higher when my monitor is only 60hz.