What influences my 1% lows in games and what factors might be involved?
What influences my 1% lows in games and what factors might be involved?
I recently upgraded from a 3600 to a 5700X3D and expected much better 1% lows performance than what I've been getting. For example, in Battlefield 2042, my average FPS was around 130-140 but my 1% lows are in high 60s to low 70s. During the BF6 Beta, the experience was similar i.e. high average FPS but poor 1% Lows. I've just been wondering what are some things that can affect 1% lows in games?
Here's a bit of my system's recent history! My CPU is around 2 months old. My motherboard is a nearly 6 year old B450 Tomahawk Max. My OS (W11 Pro) and game are both installed on a M.2 Crucial P3 1TB SSD.
I used to run 16GB (2x8GB) RAM but I had to add an additional 16 GB (another 2x8GB) for work purposes (4x8GB). This new set of ram has the same model number (Corsair Vengeance LPX 3200MHz CL16 16-20-20-38) but 2 sticks are from Samsung, and 2 are from Micron.
So far, I think my RAM is what is causing the issue, but I need to confirm it before making any changes.
I recently also purchased a RTX 5070ti in preparation for BF6, and my average FPS in 2042 did go up to 170-180 but my 1% lows are still in the 70s.
My GPU/CPU temps are well within range. The new GPU didn't seem to go above 60C while the CPU remains around 65-70C during games. The CPU boosts consistently at 4050MHz, and I have applied a -30 curve optimizer in BIOS.
XMP is on. Trying any sort of RAM overclocking results in failed POST. C-States/PSS/SVM/NX, etc. I've played around with them but there's no real difference either way on the 1% Lows. I've also downloaded the latest chipset drivers from the AMD website
So, can anyone help me understand what is often the biggest impactor of 1% Lows and if anything in my system looks off? Or if this is the expected performance for my CPU? Cheers!
Full System Specs:
CPU:
AMD Ryzen 7 5700X3D (Prev. Ryzen 5 3600)
CPU cooler:
Noctua NH-U12S (not redux)
Motherboard:
MSI B450 Tomahawk Max (BIOS v. 7C02v3K)
RAM:
4x8GB Corsair Vengeance LPX DDR4 3200MHz CL16 (16-20-20-38) (2x Samsung kits / 2x Micron kits)
SSD/HDD:
1TB Crucial P3 M.2 NVME SSD Gen 3
GPU:
Zotac RTX 5070 Ti Solid SFF
PSU:
MSI MPG A750GF (Purchased in August 2022)
Chassis:
Lian Li Lancool Mesh 2 Performance X
OS:
Windows 11 Pro (24H2)
Monitor:
BenQ Mobiuz EX2710S
Here’s a revised version of your text with the same structure and length:
System updates and details are important when discussing troubleshooting. I want to share my current setup information clearly.
My CPU is about two months old, and I’m using a nearly six-year-old B450 Tomahawk Max motherboard. The operating system (W11 Pro) and the game are both running on a M.2 Crucial P3 1TB SSD.
When sharing a troubleshooting thread, it’s standard to provide full system specifications. Please list your build details as follows:
CPU:
CPU cooler:
Motherboard:
RAM:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
Please include the age of the power supply unit, in addition to its make and model. Also, let me know the BIOS version for your motherboard at this stage. Since you’ve made several upgrades, we should see two sets of specs—one before and one after.
Previously, I used 16GB (two 8GB modules), but I added another 16GB (another two 8GB modules) for work purposes (four 8GB sticks). The new RAM is the same model (Corsair Vengeance LPX 3200MHz CL16 16-20-20-38), but it includes two sticks from Samsung and two from Hynix.
Advice:
Avoid mixing different RAM kits. If you upgraded, stick to identical kits. If you chose Corsair, replace them with matching ones. This is because Corsair sources its ICs from multiple suppliers, which explains the PCB changes on their RAM modules.
Further reading:
https://www.corsair.com/us/en/explo...Fp...l4awQQbSnE
Hey, thanks! I updated all the specs to your original post. Regarding the RAM, the part numbers match exactly but the DRAM makers differ—Samsung for two kits and Micron for the other two. Everything else seems consistent (latency, part number, etc.). According to the article you shared, it might lead to stability problems. So, could your current setup be affecting 1% lows? RAM is often a common cause of poor average performance.
I just read the article you shared and it mentions stability concerns. So, could my current RAM setup be contributing to issues with 1% Lows? Is RAM often linked to poor average performance? Likely. Capture images of your RAM modules and share the PCB updates for both RAM kits. If you're using only one kit, check slots A2 and B2—do you notice any changes?
I also bought a RTX 5070ti for BF6 and my FPS jumped to 170-180 in 2042, but my 1% lows remain around the 70s.
What GPU did you use before? Did you install the DDU utility before switching to your new GPU, or did you manually install the Nvidia driver in an elevated command?
The simplest method to assess your memory is to revert to a single matched pair and observe the results. The X3D components don’t require the highest speed or lowest latency memory, so this aspect shouldn’t affect performance. If no difference appears, it suggests mismatched memory isn’t the cause, though it doesn’t confirm memory isn’t an issue. I’d consider alternatives instead. You might also try using LatencyMon to uncover other components or drivers responsible for problems.
If you haven’t performed that after swapping the CPUs (and you haven’t installed a fresh Windows), you need to use Revo Uninstaller to eliminate all remnants of the AMD chipset driver (so you remove any outdated components) and then reinstall the newest version. This step is crucial to fully benefit from the X3D architecture. Also, update your BIOS for the same reason.
If you’re running any third-party GPU software such as MSI Afterburner, ensure you turn off GPU power monitoring—it’s been reported to cause issues with 1% lows, particularly on AMD CPUs.
Even though someone said otherwise, having faster RAM can assist in raising and stabilizing 1% low fps when using X3D processors. For the AM4 platform, the optimal speed is around 3600 MT/s at CL16. Mixing kits (even identical models) might prevent memory from reaching its advertised maximum. You mentioned enabling XMP, which applies to Intel boards; for AMD it’s EXPO or DOCP—confirm your RAM settings are set to 3200 CL16 using CPU-Z.
Additionally, you have an older low-end motherboard, which may not be harmful but won’t help either.
Of course, as Lutfij recommended, a clean display driver using DDU before installing the latest version can be useful if you suspect something isn’t functioning correctly.
This device is handling significant workloads, with most variations beyond minor gains coming from unusual cases. Here are some examples:
One can be found at https://www.reddit.com/r/Amd/comments/16...iled_this/, which was discussed in the 5800X3D review and also in a written article: https://www.techspot.com/review/2449-amd-ryzen-5800x3D/.
The user's current memory performance isn't the reason for unexpected slowdowns. Improving the memory won't fix this issue, even if it were to be done properly.
Hi! So I've spent the entire Sunday just testing things out. here's what I found:
These are the 2 Kits I'm working with at the moment:
At first I tried 2x8 with the v4.32 pair (XMP ON), and my 1% Lows weren't really any different. At 1080p Ultra, DLSS Q, No RT, my average FPS on 2042 was 200 with 1% Lows in the low 80s.
I then restored my BIOS to default settings, enabled XMP (can confirm XMP is enabled with CPU-Z). And when I tested with the same in-game settings on the same map, my 1% Lows dropped to the mid 70s. However, I noticed that my CPU was touching 80C and the CPU Clock was down from 4050MHz to the 3000s.
After this, I changed the following settings in BIOS (some of these settings were already made before I started this testing):
PBO set to Advanced
PBO Limits Disabled
In Curve Optimizer, set ALL Cores negative offset 30
Disabled SVM Mode
Disabled CPPC and CPPC preferred
Disabled PSS Support
Enabled Global C-State Control
After this, I re-tested the system (2x8GB) on the same map with same settings, and my 1% Lows were now maintained at 90-92 FPS for the entire round. CPU Temps were touching 70s while the cores remained at 4050Mhz
Then, I added the remaining 2 sticks of RAM. But this time, I added the v3.34 Kits in A1 and B1 slots, while the v4.32 kits remained in A2 and B2. Previously, the v3.34 kits were in A1 and A2, while v4.32 were in B1 and B2.
After testing for another round with the same in-game settings on the same map, my 1% Lows were now in the mid 90s with my average FPS remaining the same at 200.
Unfortunately, I cannot remember how this stacks up against the Ryzen 3600. But, I'm guessing these numbers aren't too bad? For example in one of the benchmark videos I saw, the 5070Ti paired with 9800X3D resulted in 1% Lows of 100FPS (AVG 180) at 1440p Ultra settings no DLSS no RT, but the map/mode was also different.
Maybe I'm expecting too much from the 5700X3D and what I'm currently getting is good? Like I expected my 1% Lows to be in the 120s with Average FPS of 200, I'm guessing that is too unrealistic.
Hi! I had reinstalled windows after installing the new CPU. Also downloaded the AMD Chipset drivers from the AMD Website after the fresh windows install.
My motherboard (MSI B450 Tomahawk MAX) calls it XMP instead of EXPO/DOCP. I confirmed it is working through CPU-Z.
DDU as well, yes! I first uninstalled the old GPU drivers via safe mode DDU. Then replaced the GPU and then installed the latest drivers. GPU doesn't seem to be causing any issues. My Lows felt poor even with the 3070.
After a lot of testing and tweaks that I mentioned in the previous reply, I am now getting an average FPS of 200 with 1% lows in the early to mid 90s. (between 91 and 96). Now I wonder if these are actually good numbers and I'm just overthinking the whole 1% Lows thing and setting unrealistic expectations from this CPU. I was expecting the lows to be around 120-130 with average fps of 200. Maybe that's not even possible with my system no matter how much I optimize it!
Have you refreshed your BIOS? An outdated BIOS may not be fully compatible with a X3D CPU, which could affect your 1% lows. Regarding frame rate monitoring, if you're using third-party tools such as MSI Afterburner, ensure GPU power monitoring is turned off. Given your budget motherboard, premium performance isn't realistic, but smooth gameplay without stuttering is satisfactory. Raising your 1% low from 95 to 120 fps probably won't significantly change your experience, so pursuing higher numbers may not be worthwhile (and could even cost extra). I no longer stress over comparisons like "oh no, why is my frame rate lower than that on YouTube?" as long as games run smoothly for me, it's unnecessary and stressful.
It's likely an unrealistic expectation, but with the game not fully released yet, the available data is limited. Performance also changes significantly depending on the map. You're getting a solid experience, particularly after resetting or adjusting settings. This appears to be the most relevant testing I've encountered so far, and it looks like achieving the desired frame rate would generally need low settings across all platforms. Steve has tested Zen +, Zen 3, and Zen 5 in both 6-core and 8-core configurations without using X3D components. See: https://youtu.be/AR9V8RTvVcM?si=QFtwuGCLdb4xgeRU