See, I can’t figure out why my CineBench score is 100+ less than average, AND, I just realized that even a Gears of War 5 benchmark showing someone else’s results with the same CPU GPU combination revealed I’m getting a bit less CPU framerate also, which leads me to believe it’s a CPU issue? Could my motherboard be holding it back?That looks pretty good. Hmmmm....
Did you and your wonderful friend notice at which frequency CPU peaks during tests and play? What are highest temperatures ? How much and what speed RAM ? How much resources are used at idle, for play and tests ?Hello there, ladies and gentlemen.
A few weeks ago a terrific friend of mine went through the trouble of building me my first custom computer system.
Ryzen 7 3700x
Gigabyte 2070 Super Gaming OC 3X
ASRock B450m a/c
In spite of his best efforts, as he is very experienced and a seasoned professional, I have been battling through some perceived setbacks about which I’ve bothered him to no avail.
I actually inquired about this on this forum as well as Reddit at that time, and you all were very helpful.
Since then some headway was made, but there are still a few concerns with which I’ve been struggling, so here I am.
I happened to just now put out a feeler on Reddit in reference to this, so I thought I’d save you and I the time and use an excerpt from that post here, to explain my issue(s).
“Having a bit of a head-scratching time trying to figure out why my current setup is slacking a bit behind others’ in terms of FPS and performance markers. I’ve been tinkering with synthetic benchmarks and what not to see if I’m crazy, and I can confirm to some degree that something seems to be holding back. The wonderful soul who built my system for me I have beyond exhausted asking similar questions to this, so I thought I’d take it easy and inquire elsewhere for now.
Most times with Battlefield V and Warzone, I’m getting around 100+ FPS... seems about right from what I’ve seen for this combo. That wasn’t the case for a bit as I had to do some experimenting and find out what was going on... At first I was experiencing stutters here and there, very small, very short, very random, but that has since evened out... it seems though, that in Gears of War, among some other games, in the benchmark, others with the same/similar setup, same GPU nonetheless, are getting 5-10fps average more than me, sometimes more, identical settings, especially in the CPU FPS portion of that benchmark. I notice I score a bit above the modern gaming PC mark and below the High End gaming PC mark for TimeSpy and what not, and in PassMark 3D mark my card scores in the 96th percentile. Something I do notice - the MSI afterburner (no I do not currently have an OC going) overlay often signals the power limit indicator. What that means, I don’t know. Does it make a difference if Gears is running from the Xbox game pass app vs. Steam?
As I said, I’m new. I have little idea how to compare my experience to similar systems, though I’ve been doing that ad absurdem for the last two weeks.
I’ve done all the driver/bios/chipset reinstalling and updating, XMP is ON, game bar is off, game mode is off... etc...
I wonder if it’s my motherboard that could be holding it all back. Everybody else seems to have an X570. Could that be the key to my performance?
Could it be a CPU thing? I actually just swapped my GPU for the same model with a higher clock, so I don’t know that my GPU is the weak link here.
My CPU scores around a 4750-4790 in Cinebench R20 and something around 30-80 points on average below what I see for my system’s components on Geekbench.
Perhaps I just have hit the mid range of the silicon lottery, for the CPU at least. Maybe there is a rogue setting I have yet to change somewhere... Perhaps I’m deluded!
Would love to hear your thoughts.”
I know that together, we can solve this! Thank you in advance for your expertise and time. I’d be happy to do any tests and benchmarks, should you need more information.
So! Itemized responses:Did you and your wonderful friend notice at which frequency CPU peaks during tests and play? What are highest temperatures ? How much and what speed RAM ? How much resources are used at idle, for play and tests ?
If CPU was running at it's peek, CB r20 scores should be about 5000.
Looks like it's "slacking" by about 100MHz. is PBO enabled, CPU temps under 70c under full load ? Ryzen are very sensitive to temps, best at 62-65c and dropping fast above 70c.So! Itemized responses:
Looks like it’s different for different games. Some peak at 4300 or so. Highest temps can be around 82, but that is quite seldom. RAM is 2x8GB for a total of 16GB at 3200Mhz XMP on.
Cinebench R20 usually tops out at around the high 4700’s. Since my previous post and this, a big change I made was deleting the antivirus program installed initially: Sophos Home. Still, even after doing these tests with nearly every background program and processes meticulously combed through and ended after a fresh boot, we get no more than 4790. Could it be my CPU just slacking behind a bit?
This is a very fair observation.
PBO is not enabled, however I’d imagine that I should be able to achieve a certain level of performance at stock speeds, no? The scores to which I aspire are simply average results for my CPU and GPU on said tests such as FireStrike, TimeSpy, Cinebench, and Geekbench. Ironically in Heaven Extreme, it scores higher than most have expressed with my components...Looks like it's "slacking" by about 100MHz. is PBO enabled, CPU temps under 70c under full load ? Ryzen are very sensitive to temps, best at 62-65c and dropping fast above 70c.
Yup, I am pretty sure I will have said something very similar to this in that other thread too.
Cinebench, and Geekbench are more system wide benchmarks while others specially Heaven Extreme are more graphic oriented.PBO is not enabled, however I’d imagine that I should be able to achieve a certain level of performance at stock speeds, no? The scores to which I aspire are simply average results for my CPU and GPU on said tests such as FireStrike, TimeSpy, Cinebench, and Geekbench. Ironically in Heaven Extreme, it scores higher than most have expressed with my components...
I genuinely appreciate your most thoughtful approach.Yup, I am pretty sure I will have said something very similar to this in that other thread too.
You simply cannot build a system and then go comparing it against YT reviewer videos. It's NOT an apples to apples comparison. You have no idea exactly what is on/off, resolutions, timings, overclocks. Almost every single reviewer out there is being paid and supplied parts in return for positive reviews. It is VERY seldom that anyone will break that status quo.
The fellow that runs Tech Yes City actually told on himself as well as MANY other reviewers in a video a few weeks ago. He flat out stated that his intention is selling computers and in light of that certain allowances are made. In particular one of the main tools he uses in that respect is lower resolution and scaling. Other settings will certainly be 'off stock' as well.
As a final thought, if your system is running well otherwise, not restarting, not overheating, etc. It seems you are getting playable frames and such. Enjoy the build for what it is.
So then; the CPU or may be to blame here.Cinebench, and Geekbench are more system wide benchmarks while others specially Heaven Extreme are more graphic oriented.
That's what some benchmarks are good for, to see if you are getting your moneys worth. It's a controlled environment which loads CPU to it's(near) maximum and keeps it thereSo then; the CPU or may be to blame here.
I do not rule out the distinct possibility that there could be a setting or program that I haven’t switched on or off; as I said I am very new to this vast universe!
Would you mind guiding me as to how to limit temperatures? In addition, I’m assuming this is all done in Ryzen Master.That's what some benchmarks are good for, to see if you are getting your moneys worth. It's a controlled environment which loads CPU to it's(near) maximum and keeps it there
although your games may or may not push it that far. CB score prompted me to ask you if PBO was enabled as I knew it most probably wasn't. Telling me some game scores would mean nothing if I didn't have it to compare with.
In short, if you want to get everything as good as possible. PBO needs to be enabled and set and temps no more than 70c, preferably 62-65c. Every setup working at it's peek under those conditions, will get very similar results in CB r20 for instance. It's not some competition thing but just a way to see if everything at it's best. 5000 points +/- 20 is what I see in most systems with 3700x. If you let CB run 3-5 times in a row, it will drive temps to their maximum.
It was proven time and again that Ryzen is sensitive to temperatures, in most tests including mine 62-65c is about maximum at which PBO will get highest frequencies. In addition, last batches (end last year and up) are somewhat better than earlier ones. That results with 50 to 100MHz better top frequency.Would you mind guiding me as to how to limit temperatures? In addition, I’m assuming this is all done in Ryzen Master.
Prepare for a question riddled with ignorance:
Is any of this potentially harmful to my build? I’ve looked around about overclocking and have used afterburner a few times here and there, CAREFULLY and modestly, but I understand this is a touchy part of the optimization process.
Additionally, the Gears benchmark is actually a bona-fide benchmark scene which is built into the game, and gives a bunch of information at the end.
I made sure to set my settings in accordance to what others had in their videos of the benchmark process, and it seemed I scored 10-15 avg FPS lower, with the CPU framerate being the most noticeable difference.
One example is an unpopular video of a build with my same CPU and GPU, settings with which I score 69-71 average FPS.View: https://youtu.be/tixo0VADtZc
Aside from that....
Here are a few examples (I don’t expect you to view all of these, but have at it if you are so inclined...)
This test shows the 2070 Super with the Gears benchmark at 3 different settings. We will observe the 1440p benchmark, where they achieve an avg FPS of 75. If I set all my settings equal to those (though there are some like MSAA that don’t show there, although when someone says ultra they usually mean, in this game, using the ultra setting preset which I believe sets everything to 16x) I get about 69-71.
A similar test, again, at the 1440p portion, showing the 2070 Super at ultra settings reaching an average FPS of 81. Some might wonder if there is an overclock in place for the card, but look at the clock speed! <1900mhz?
Exhibit C (lastly):
Again, only referring to the beginning of the video with the 1440p portion. An average of 81fps. COULD it mean anything that the 1440p portion is downscaled from a native 2160p? Hmm.
Interesting. Most of my gaming is done at 1440p. I don't think I've used this for anything at 1080p since it was put together.It was proven time and again that Ryzen is sensitive to temperatures, in most tests including mine 62-65c is about maximum at which PBO will get highest frequencies. In addition, last batches (end last year and up) are somewhat better than earlier ones. That results with 50 to 100MHz better top frequency.
Just to have clear known example let me compare my (first batch) 3700x at CB R20.
With my setup CB scores at up to 65c are about 5000 with peek frequency 4.39GHz at one core at a time and 4.25 on others. at 70c score is down to 4910 or thereabouts , frequency peek and average on other cores by about 50-70MHZ, at 75c, frequency drops by another 100MHz across the board,. dropping CB score to 4800+.
So much about CPU. With gaming up to 1080p and normal frequencies, it's CPU that does most work, above those GPU gets more importance as well as game settings.
In my experience, overclocking GPU doesn't bring much practical boost unless you are just at border of usability and OC may help jump over that hump. I wouldn't dwell to much on GPU benchmarks unless like that one is built right into it or it's just for particular game. Just too many variables.