• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Ryzen Owners Zen Garden

What did you find so hard working on it?
I thought it was pretty straightforward to work with.
The way all of my cables were tucked away, a lot of it was relaying pathways.
 
What did you find so hard working on it?
I thought it was pretty straightforward to work with.

The Torrent Compact manages to be kinda weird w.r.t. cable management. It's an extremely similar layout to the Define 7 Compact, yet it was much less trying to route that build than either of the two TC's I've done. Love it otherwise, though I recommend blocking off the bottom grille somehow if doing a pure front --> back airflow strategy. Netted me a ~3C drop on CPU and GPU.
 
Yup, it takes a little planning. Otherwise, great case.
 
Odd, didn't have any issues with cable routing.
 
Odd, didn't have any issues with cable routing.
Me neither. But cueing up YouTube, partaking in Cannabis consumption, and just generally taking my time.. yes I could have done it in half the time because I am fairly good at what I do.
 
My cable routing is simple: keep the front neat, tuck everything in the back. Sometimes I regret it, though, especially today, installing a new hard drive in the cage next to the PSU. :D
 
Odd, didn't have any issues with cable routing.

It could very well be a me problem. All I know is that both TCs took me twice the time of the D7C.
 
My cable routing is simple: keep the front neat, tuck everything in the back. Sometimes I regret it, though, especially today, installing a new hard drive in the cage next to the PSU. :D
I know I just balled up the wires for the 5 rad fans and jammed them into a corner.. now I have to dig it all out again :laugh:
 
It could very well be a me problem. All I know is that both TCs took me twice the time of the D7C.
Well, I guess I don't have that many cables either. No SATA drives helps reduce the clutter.
 
For me the source of back panel woes was always the same culprit, the friggin' honey boo-boo sized 24-pin cable.
No issues w my Meshify II Compact, but it was always a problem before that especially w my Arc Mini.
 
Well, I guess I don't have that many cables either. No SATA drives helps reduce the clutter.
Mine is fully loaded, I have 3 drives in the back that are not listed as well.
 
I know I just balled up the wires for the 5 rad fans and jammed them into a corner.. now I have to dig it all out again :laugh:

That's my standard procedure as well. I quite like the building process, but cable routing is like pulling wisdom teeth. Sufficient space in the rear has always been one of the main points when choosing a case.
 
Mine is fully loaded, I have 3 drives in the back that are not listed as well.
Yeah, that would make a bit of a harder build.
The case needed an extra 5-10 mm of space between the side plate and the motherboard.
Haven't had a system this quiet in years though and that includes cases with sound dampening, that I had to open up the top cover on when gaming, as they just got too hot.
 
Yeah, that would make a bit of a harder build.
The case needed an extra 5-10 mm of space between the side plate and the motherboard.
Haven't had a system this quiet in years though and that includes cases with sound dampening, that I had to open up the top cover on when gaming, as they just got too hot.
Love this case :)
 
User of a 5800x3D since 10th March, 81c max temp under the Thermalright PA cooler.
My biggest issues came from the GPU and the way that the Corsair 4000D does not shift air fast enough to stop warm air build up, it is tackled now thankfully.

Re-pasting the TUF ASUS 7900 XT dropped temps by some 15c.... yes you heard right at the same comparable fanspeed, unlocking whole new overclock abilities.

The CPU I do wonder if it would benefit from Prolimatech PK-3 like the GPU did but 81c I am happy with.

From the TPU review of the XTX version with the same TUF cooler it was noted for it's amazing OC headroom, yes my XT version cannow easily sustain 3.2ghz and remain daily usable as in sound output.


"While the AMD reference RX 7900 XTX was 8% faster than the RTX 4080, the stock ASUS TUF is 11.5% faster, if you add OC+UV on top, you're getting 23.1%—this looks much closer now to what AMD has promised us. "


So far the only game to give me troubles is Forza Motorsport which is a mess for most people.

And currently the PC is airflow optimized so Window>>>>front intakes>>>>CPU / GPU>>>>>2 exhausts.

IMG_20230904_205159.jpg
IMG_20231017_141806.jpg
 
The CPU I do wonder if it would benefit from Prolimatech PK-3 like the GPU did but 81c I am happy with.
No. As long as it doesn't reach the 89 °C limit, it will use its maximum boost.

Nice system, by the way. :)
 
Your settings are all over the place. Your tRFC is insanely high if you have Hynix chips.
Those were automatic settings from XMP/EXPO - I'll look into your suggestions when i boot it up again next.

If the boards are applying crap defaults, that's why some users have worse experiences than others.

I guess that confirms it... F is in no sync to anything, and only M and U matter?
Definitely isn't a flat sync between them, In fact it seems F needs to be less than U.
The BIOS on my Taichi Carrara does specify that, and that may be why people see negative scaling in some situations but misinterpreting why.
U needs to be below F, but the amount isn't specified.

This random reddit quote says ~3 notches - but they're talking F and M here.
1698119371135.png


1698119489599.png


3000/2000/1500 may be easiest to run with no slow boots for many people and minimal performance loss (Needs testing) - XMP varies wildly on boards here, some run the half U clock, some dont even push to 2000 on the Fclk.

I get the feeling these values aren't perfectly mathed and are just a representation and the actual dividers behind the scenes are the reason for the contradictory results - like that mystery NB clock in Aida64 that bounces around like mad

'm working on the theories for WTAF is happening and why the brands all handle it so differently, so i have a plan of attack when i test it next.


I dont think it was possible to know this before the new BIOS updates allowed ~DDR5 8000, my testing with the DDR5 8000 kit has shown that XMP 7600 with loose timings has less latency than XMP 6000 on my system - and that shouldnt happen
 
Definitely isn't a flat sync between them, In fact it seems F needs to be less than U.
The BIOS on my Taichi Carrara does specify that, and that may be why people see negative scaling in some situations but misinterpreting why.
U needs to be below F, but the amount isn't specified.

This random reddit quote says ~3 notches - but they're talking F and M here.
View attachment 318699

View attachment 318700

3000/2000/1500 may be easiest to run with no slow boots for many people and minimal performance loss (Needs testing) - XMP varies wildly on boards here, some run the half U clock, some dont even push to 2000 on the Fclk.

I get the feeling these values aren't perfectly mathed and are just a representation and the actual dividers behind the scenes are the reason for the contradictory results - like that mystery NB clock in Aida64 that bounces around like mad

'm working on the theories for WTAF is happening and why the brands all handle it so differently, so i have a plan of attack when i test it next.


I dont think it was possible to know this before the new BIOS updates allowed ~DDR5 8000, my testing with the DDR5 8000 kit has shown that XMP 7600 with loose timings has less latency than XMP 6000 on my system - and that shouldnt happen
Personally, I think that reddit post is just as much an exact science as tea cup fortune telling is. If there really was some ratio on F, then the BIOS would ask for the ratio, not the actual clock.

Even if there is some latency penalty on running F out of ratio, there's no way of knowing in real-world programs.
 
Personally, I think that reddit post is just as much an exact science as tea cup fortune telling is. If there really was some ratio on F, then the BIOS would ask for the ratio, not the actual clock.

Even if there is some latency penalty on running F out of ratio, there's no way of knowing in real-world programs.
I've seen the same things with what i've tested here, which is why it has some merit
Theres some oddities with the settings, beyond the simple 1:1 of AM4, that's for sure.


And as one obvious example, borderlands 3 was insanely sensitive to memory latency - that's why the x3D chips blasted performance sky high in that as a benchmark. Memory latecy doesn't matter for low FPS gameplay, but it really does in higher FPS usage - and that's an issue when benchmarks are usually done at ultra settings, where actual gamers try to match their refresh rates
 
I've seen the same things with what i've tested here, which is why it has some merit
Theres some oddities with the settings, beyond the simple 1:1 of AM4, that's for sure.


And as one obvious example, borderlands 3 was insanely sensitive to memory latency - that's why the x3D chips blasted performance sky high in that as a benchmark. Memory latecy doesn't matter for low FPS gameplay, but it really does in higher FPS usage - and that's an issue when benchmarks are usually done at ultra settings, where actual gamers try to match their refresh rates
Ideally, everyone is aiming for Ultra, in my opinion. The only time that changes is when your PC can't run it.

If by "gamers" you mean e-sport competitors, then that's a different story and a world I don't know much about.
 
Ideally, everyone is aiming for Ultra, in my opinion. The only time that changes is when your PC can't run it.

If by "gamers" you mean e-sport competitors, then that's a different story and a world I don't know much about.
In reality, i dont know a single person that does that in the all the years i've ran LAN parties of hundreds of people, as well as the ones i game with daily.

Unless they're outright old games that are easy to run, people target a framerate and not the settings - those of us with high refresh rate displays absolutely turn settings down til we get it, as we'd rather lose film grain and motion blur than use DLSS/FSR.

Settings like those are enabled in ultra, yet i dont know a single person that actually wants motion blur enabled.


It's irrelevant as a user with a 3090 on a mix of ultra and medium settings could get the same performance as a 4090 on ultra, meaning the CPU and RAM performance end up mattering equally as much - lowering settings just gives you the equivalent of a faster GPU, as far as the rest of the system is concerned. More FPS? More work to do.
 
Last edited:
In reality, i dont know a single person that does that in the all the years i've ran LAN parties of hundreds of people, as well as the ones i game with daily.
That's weird. I don't know a single person who doesn't do that.

Unless they're outright old games that are easy to run, people target a framerate and not the settings - those of us with high refresh rate displays absolutely turn settings down til we get it, as we'd rather lose film grain and motion blur than use DLSS/FSR.
That's what VRR exists for, I guess. I see 70-ish FPS just as smooth as 144.

Graphics are part of a game, imo. Turning it into a bland smear just for the kills makes your experience a bit more bland as well. But each to their own. :)

Settings like those are enabled in ultra, yet i dont know a single person that actually wants motion blur enabled.
I never really wanted it in any game, but if it's there, it doesn't bother me.
 
Graphics are part of a game, imo. Turning it into a bland smear just for the kills makes your experience a bit more bland as well. But each to their own. :)
Have you never seen games that ultra turns it INTO the bland smear?

You game with motion blur, film grain and all those extras enabled at all times? Path tracing in CP2077, no matter what?
 
Have you never seen games that ultra turns it INTO the bland smear?

You game with motion blur, film grain and all those extras enabled at all times? Path tracing in CP2077, no matter what?
Motion blur doesn't bother me if my frame rate is decent. Film grain is pointless, though, so I turn it off. Path tracing looks cool, I guess, but it's up to whether your GPU can handle it or not.

But I don't turn settings off just to get a bazillion FPS.
 
Tested the U/F/M clock settings on AM5 a bit better
Theres definitely some hidden math at work, and that makes it very hard to tell people what clocks to use.

Theres nothing between 1800 and 2000, so it's 1800/2000/2033/2066/2100+

That said, the loss from going *TOO HIGH* is far worse than just staying at 1800 sometimes


Focus on the Memory latency only here, as that's what is being tested

Fclk: 2133 = 118.5 ns (Memory read and write tanked massively here)
Fclk: 2066 = 72.9 ns
Fclk: 2033 = 71.8 ns
Fclk: 2000 = 72 ns
Fclk: 1800 =72.7 ns

1800 gave a lower or equal result to 2066 in many of the test runs


For DDR5 6000, the answer is simple:

2033 is the best, 2000 is second best, and 1800 is perfectly fine - especially if it resolves slow boot/cold boot issues.
Going for maximum speeds above 2033 is very likely NOT worth it, unless your RAM speed is higher too.

Screenshot 2023-10-30 134007.png
Screenshot 2023-10-30 141653.png
Screenshot 2023-10-30 145107.png
Screenshot 2023-10-30 144607.png
Screenshot 2023-10-30 144018.png


This is whats thrown me off with a lot of what people have been suggesting, because there are far too many people saying "clock as high as you can go" and the performance loss from doing so goes from 'trivial' to "nailgun to the testicles" very very easily
With the slow boot issues some people have, is shaving 1ns from 1800 to 2000 worth it on those systems? My vote goes to definitely not. Tune in lower voltages and timings instead.


I'll test this with 7600MT/s soon, to see if it's just 2033 being the best value or if it maths with the RAM speed

But I don't turn settings off just to get a bazillion FPS.
I'm only talking about turning things down to stay within your refresh rate, 60Hz users wont have to do anything here, but 144 and up definitely will, no matter the harwdare.
My brother with his Gsync ultimate display and its support of 1Hz means he can handle anything smoothly, while Freesync 2 displays with a 48Hz minimum are entirely different

I could word it better as "I don't know anyone who'd run maximum settings rather than an acceptable frame rate" (With acceptable coming down to the display, usually)
 
Last edited:
Back
Top