Low FPS

[BattleToads]

OK so it seems Rainbow Six Siege is running ok. 144 FPS on ultra most of the time. CPU at 50%, GPU at 60-70%. Some stuttering every now and then, but it could just be internet lag. Street Fighter 5, however, (my main competitive game) is running at 55 FPS. I can feel it in the game and it is imperative that it run at 60 due to the mechanics of the game. Not sure what's going on here. 

[BattleToads]

Just tried Far Cry 6. Typical BS. Took FOREVER to load the game. (PSO cache countdown) FPS in benchmark ran even worse on high than ultra. FPS will read above 60 but still looks and feels extremely choppy. Something is severely wrong. This is two SSDs that this has occurred with. Is it the motherboard? Is it the CPU? I need help here. I have never had problems like this. 

[BattleToads]

I feel like I should be getting a lot better than this... I just remembered. Doesn't the 1600AF have a 1st gen memory controller? And wasn't the first gen memory controller kinda finnicky? I am using one stick of G. Skill Ripjaw and one stick of G. Skill Aegis with the same timings. According to the BIOS, one is made by Samsung, one is made by Hynix. I replaced the other Ripjaw because it was throwing out errors in memtest. I did spill water on this PC two years ago and fried the GPU, but I don't know if this has anything to do with it. I've been running it fine other than the FPS/stuttering issue. I really don't know at this point. 

[lordmogul]

Even with that less awesome memory controller of early Ryzen (compared to the later ones) performance should be better than that.

Since you first mentioned Doom 2016, I play the game on an older and less powerful system (i5-3570K + GTX 1060) and from a mechanical drive and get better performance than that. I even ran it on my old Core 2 system out of curiosity and was surprised how well it did. So you shouldn't have any issues with that, even if you don't run the game from SSD.

 

something else has to be off here. Try running the usual benchmarks (3DMark, Cinebench, any game with an ingame benchmark sequence etc) and compare your score to other systems with the same CPU/GPU, you should be maybe 5-10% towards the majority of the systems listed.

They differ a bit because of different boost behaviour and overclocking and such, but generally should be in a similar range. (around 7000-7800 in TimeSpy, 175,000-190,000 in Firestrike, around 6700-6900 in Cinebench R23 and 2600-2700 in R20, just for a rough estimate)

[BattleToads]

1 hour ago, lordmogul said:

Even with that less awesome memory controller of early Ryzen (compared to the later ones) performance should be better than that.

Since you first mentioned Doom 2016, I play the game on an older and less powerful system (i5-3570K + GTX 1060) and from a mechanical drive and get better performance than that. I even ran it on my old Core 2 system out of curiosity and was surprised how well it did. So you shouldn't have any issues with that, even if you don't run the game from SSD.

 

something else has to be off here. Try running the usual benchmarks (3DMark, Cinebench, any game with an ingame benchmark sequence etc) and compare your score to other systems with the same CPU/GPU, you should be maybe 5-10% towards the majority of the systems listed.

They differ a bit because of different boost behaviour and overclocking and such, but generally should be in a similar range. (around 7000-7800 in TimeSpy, 175,000-190,000 in Firestrike, around 6700-6900 in Cinebench R23 and 2600-2700 in R20, just for a rough estimate)

Ran The FarCry 6 Benchmark and it was painful to watch. 50 FPS and stutter-city. You could watch the framerate take hard drops to zero every few seconds, otherwise running between 50-70 FPS. Performance was somehow worse on high than on ultra. Even on the cutscenes, the game looked horrible with the stutter and screen tear. 

[BattleToads]

Would this have something do do with BIOS update/chipset update? I'm about to throw this god damned thing out the window. 

[lordmogul]

Just a guess, but what drive do you have the games installed on, and is there anything else working on them? I'm thinking about issues with loading stuff from the drive, because some other program, or even Windows itself, is hammering it, leaving no bandwith for the game.

[BattleToads]

18 minutes ago, lordmogul said:

Just a guess, but what drive do you have the games installed on, and is there anything else working on them? I'm thinking about issues with loading stuff from the drive, because some other program, or even Windows itself, is hammering it, leaving no bandwith for the game.

FireCuda 520 - only drive in the system

[Guest]

On 6/17/2022 at 5:38 PM, BattleToads said:

Ryzen 5 1600 AF

RX 6600 8GB

16 GB CL16 3200 

 

I can't even get 60 FPS constant in Doom 2016 on medium... Constant frame drops into the 30s and 40s. CPU and GPU are nowhere near capacity yet performance is horrible. Open GL 4.5. 

 

What's going on here? 

This processor has lower IPC than a 6700K. 

 

It's not how many cores the game uses, but the horse power per core. 

 

Upgrade the cpu, I bet the gaming experience will become much much better. 

 

That is all.

[BattleToads]

4 minutes ago, Guest 5150 said:

This processor has lower IPC than a 6700K. 

 

It's not how many cores the game uses, but the horse power per core. 

 

Upgrade the cpu, I bet the gaming experience will become much much better. 

 

That is all.

Then why am I seeing YT videos of people with this CPU getting much higher FPS?

[Guest]

36 minutes ago, BattleToads said:

Then why am I seeing YT videos of people with this CPU getting much higher FPS?

I didn't base my reply on YouTube comparisons. Nor should your diagnostics.

 

The 1600AF is a budget processor. A bit slower than my 2700X and a lot slower than my 8700K. Which that chip is much much slower than a budget 12400F.

 

The bottleneck to your rx 6600 is the cpu. 

 

Maybe you should be getting 10fps more, but that could be from in game settings like having post processing enabled or multi-core rendering enabled.

[BattleToads]

3 minutes ago, Guest 5150 said:

I didn't base my reply on YouTube comparisons. Nor should your diagnostics.

 

The 1600AF is a budget processor. A bit slower than my 2700X and a lot slower than my 8700K. Which that chip is much much slower than a budget 12400F.

 

The bottleneck to your rx 6600 is the cpu. 

 

Maybe you should be getting 10fps more, but that could be from in game settings like having post processing enabled or multi-core rendering enabled.

OK that would be nice if it's as simple as a CPU swap. Questions - would this also explain my slow SSD benchmarks? Since the CPU is what powers the PCI-E lanes, I suppose? Should I also get a new motherboard since mine is only supportive of up to Ryzen 3000 as is, and I'd rather not mess with flashing the BIOS. (even though I just tried, and apparently this USB drive isn't bootable) 

[Guest]

Just now, BattleToads said:

OK that would be nice if it's as simple as a CPU swap. Questions - would this also explain my slow SSD benchmarks? Since the CPU is what powers the PCI-E lanes, I suppose? Should I also get a new motherboard since mine is only supportive of up to Ryzen 3000 as is, and I'd rather not mess with flashing the BIOS. (even though I just tried, and apparently this USB drive isn't bootable) 

A faster cpu should improve most all benchmarks. It is the brains of the operation after all. But perhaps not super greatly depending on the benchmark comparison to the spec of the drive. If it's close already in read and writes, then it won't improve greatly no.

 

Yes for a new motherboard if you did like a 5600X. I think you'll be super happy with the gaming performance. It will be around 50% increased at least. The newer processor might even be able to run your memory at it's rated speed.

 

And speaking of memory, I have a feeling you're not running a very high frequency. If you can improve the memory frequency, it should improve your performance on the current build, BUT the 1600F is only good for up to 2933mhz at best. With a 5600X, 3200mhz memory or faster would further increase the performance. It would be a night and day difference.

 

Now I'm not telling you to spend money. Just giving some pointers and information. 

[Montana One-Six]

Before you buy anything. I would try to make some more tests to get a bit better understanding of the current issue because right now you only really know that something is not working as it should which can be a lot of things and randomly buying things in hopes they will fix it is a pretty expensive way of fixing things.

 

First things first I would download the 3D Mark demo from Steam and run TimeSpy as that benchmark will give you scores for CPU and GPU which can be compared to a fairly big database of other PCs. If any score seems way off investigate further in that direction.

 

On 6/25/2022 at 10:27 AM, BattleToads said:

I am using one stick of G. Skill Ripjaw and one stick of G. Skill Aegis with the same timings.

Remove the XMP profile and use the RAM at base clocks (2133) and timing and test again. Also it is probably a good idea to use both ripjaws or both aegis instead of mixing them especially if they are not using the same memory dies.

 

23 hours ago, BattleToads said:

Would this have something do do with BIOS update/chipset update?

What Motherboard do you have? Is the newest BIOS installed?

[BattleToads]

4 hours ago, Montana One-Six said:

Before you buy anything. I would try to make some more tests to get a bit better understanding of the current issue because right now you only really know that something is not working as it should which can be a lot of things and randomly buying things in hopes they will fix it is a pretty expensive way of fixing things.

 

First things first I would download the 3D Mark demo from Steam and run TimeSpy as that benchmark will give you scores for CPU and GPU which can be compared to a fairly big database of other PCs. If any score seems way off investigate further in that direction.

 

Remove the XMP profile and use the RAM at base clocks (2133) and timing and test again. Also it is probably a good idea to use both ripjaws or both aegis instead of mixing them especially if they are not using the same memory dies.

 

What Motherboard do you have? Is the newest BIOS installed?

I took my PC to a repair shop a couple years ago and they said one of the RAM sticks was spitting back errors in memtest, so I swapped it out with an Aegis stick as you can't just buy Ripjaw in a single stick, only kits. At least this was the case at the time. 

 

The mobo is a cheap Asrock B450M-HDV. The BIOS is only current to 2019. I'd have to get some bootable USB drives to update it. 

[BattleToads]

I think I may have fixed the issue. I disabled the XMP profile and games are running much smoother. When frames drop below 60 now, it feels more like a typical CPU bottleneck. Generally occurs when CPU is pulling over 50W of power. 

[Montana One-Six]

13 hours ago, BattleToads said:

I think I may have fixed the issue. I disabled the XMP profile and games are running much smoother. When frames drop below 60 now, it feels more like a typical CPU bottleneck. Generally occurs when CPU is pulling over 50W of power. 

Nice. Without the XMP you will lose some performance overall. But you could try to get some of it back by manually setting frequency and timings in BIOS if you feel like it.

 

Also a BIOS update could be beneficial to RAM oc stability and issues.

[BattleToads]

So what about SAM and 4G encoding and all that? Is that worth messing with?