Ryzen Owners Zen Garden

[Kissamies]

Apparently, Zen 4 isn't too sensitive, either:

DDR5 Ryzen 7 7700 DDR5 memory scaling review

You've likely just read our Ryzen 7000 reviews. and will have noticed that AMD calls DDR5 6000MHz /30CL a sweet spot. It, however might not be a sweet spot for your wallet, ergo we take a look at how... Performance - Gaming RTX 3090 - Assassins Creed Valhalla

www.guru3d.com

This calls for my own testing, I guess.

I guess that older Zens are more sensitive. I mean, even with 1080 Ti I saw a noticeable gain in games with faster RAM

 

[AusWolf]

Weird conclusion to come to since even in the limited selection of games they tested, there are instances where RAM makes a difference at all resolutions.

It's perfectly understandable that the games you play may not see much benefit, but blanket statements aren't very accurate.

Vcache generally seems to level the playing field, however, so RAM doesn't always make a huge difference

I just try to see the usefulness of difference. The biggest one I see is Far Cry 6 at 1080p, which is 135 vs 150 FPS between 4800 vs 6000 MHz, which still isn't significant enough to matter, in my opinion.

 

[Kissamies]

I just try to see the usefulness of difference. The biggest one I see is Far Cry 6 at 1080p, which is 135 vs 150 FPS between 4800 vs 6000 MHz, which still isn't significant enough to matter, in my opinion.

Well, 10% is still somewhat a noticeable difference even if not that significant.

 

[tabascosauz]

I guess that older Zens are more sensitive. I mean, even with 1080 Ti I saw a noticeable gain in games with faster RAM

Think I heard bz mention in passing in a recent video that dual rank for DDR5 doesn't matter as much as it did for DDR4. Something to do with having a greater number of effective bank groups.

Though, I was only half-listening while doing something else, so it might be an intel only thing, who knows.

I just try to see the usefulness of difference. The biggest one I see is Far Cry 6 at 1080p, which is 135 vs 150 FPS between 4800 vs 6000 MHz, which still isn't significant enough to matter, in my opinion.

Like I said, perfectly understandable if you don't mind it on your setup. But most people are here to get the most out of what they buy. Running JEDEC on a 6000 kit doesn't exactly fit that bill.

And if you have a 60Hz display, it's hard to justify anything, honestly. Everything feels the same anyway

 

[HD64G]

-15 all core would end up in whea 19 reboots during low power states in my case. Can't imagine going above 20. Like I said, I don't even know what would RM suggest because it never finished.

RM CO test is good only for per-core optimization. So, let it run every core seperately and adjust those one by one.

 

[AusWolf]

Like I said, perfectly understandable if you don't mind it on your setup. But most people are here to get the most out of what they buy. Running JEDEC on a 6000 kit doesn't exactly fit that bill.

And if you have a 60Hz display, it's hard to justify anything, honestly. Everything feels the same anyway

I do want to have 6000 MHz and use my system to its full potential - that's why I bought this kit. It's only that the lockup during every single cold boot is extremely annoying. I'm trying to find the tradeoff. If I have to lose 5% game performance that I don't notice anyway (I have a 60 Hz display), then so be it, I guess. Not ideal, but not bad.

If anyone has a better solution (maybe I need to manually do something with my voltages), please shout up!

 

[freeagent]

Hey! I found pictures of my rig being used in a review, how awesome is that??

Photos were shot on an iPhone XS Max lol

Made my day!

Thermalright Frost Commander 140 (xtremesystems.org)

Down in the AMD part

Edit:

I had the cooler upside down without knowing it at the time, that is why it is so close to the pcie slot. Turn it around and you can stick a couple fingers in there

But all of those Strix shots and case shots with my 120x38s and 980 Classified, those are all me

 

[AusWolf]

Hey! I found pictures of my rig being used in a review, how awesome is that??

Photos were shot on an iPhone XS Max lol

Made my day!

Thermalright Frost Commander 140 (xtremesystems.org)

Down in the AMD part

Edit:

I had the cooler upside down without knowing it at the time, that is why it is so close to the pcie slot. Turn it around and you can stick a couple fingers in there

But all of those Strix shots and case shots with my 120x38s and 980 Classified, those are all me

I would claim for unauthorized use of intellectual property if I were you.

 

[freeagent]

I would claim for unauthorized use of intellectual property if I were you.

Ehh.. I don’t mind

 

[Mussels]

It seems my 5800x3D greatly prefers my new alphacool block over my EK block

sure it hasnt been running long, but the heat would build up fast on the EK - look at that lack of spread


Although at present it looks like frodo needs to drop the one ring into it

 

[freeagent]

Nice and tight

Mine is just a little looser after 30 mins

 

[MagnyCours]

Updated my BIOS from version F6B to F7A. After the update, my previously stable memory OC settings were suddenly giving me all kinds of issues (unable to POST, freezing in BIOS, Windows memory_management BSOD, etc.)

Turns out this setting was the culprit:

It was set on "Auto" on the newest BIOS, which probably meant that it was Enabled by default. So I disabled it and all the issues I had were gone! It was probably causing all kinds of memory training issues when it was enabled.

 

[Mussels]

cross-posting this into another thread, but awww-yis
Nothing like cooling that works properly, and power efficent gaming (R23 for 30 minutes, 2 hours of 4k140 gaming and then an hour of TPU)

No more random 80C spikes, that was apparently something to do with the EK block

(That 633W spike when the FPS limiter wasn't active yet and afterburner wasnt limiting things tho, oops XD)

 

[tabascosauz]

Updated my BIOS from version F6A to F7B. After the update, my previously stable memory OC settings were suddenly giving me all kinds of issues (unable to POST, freezing in BIOS, Windows memory_management BSOD, etc.)

Turns out this setting was the culprit:

It was set on "Auto" on the newest BIOS, which probably meant that it was Enabled by default. So I disabled it and all the issues I had were gone! It was probably causing all kinds of memory training issues when it was enabled.

Is this a new AGESA feature to address long AM5 POST times? Seems like a good solution......if it worked...lol

 

[MagnyCours]

It probably is! I did notice that boots on restarts were almost instantaneous when the setting was enabled, but it was unstable as hell so.... back to 30 sec. long POST times for me.

 

[Chomiq]

I gave it a shot once, took over an hour and a half (with two restarts) only to fail at 99% which then had me restart from scratch after reboot.

Instead I stick to -15 on silver/gold cores and -10 on everything else.

And of course two days after I posted this I got WHEA reboot on Core 3 (gold). All core -10 it is.

Honestly, curve optimizer is a PITA, it can be stable for weeks and yet somehow you run into a random reboot in hour 6 of gaming session.

 

[tabascosauz]

And of course two days after I posted this I got WHEA reboot on Core 3 (gold). All core -10 it is.

Honestly, curve optimizer is a PITA, it can be stable for weeks and yet somehow you run into a random reboot in hour 6 of gaming session.

Done any proper testing with corecycler, OCCT or ycruncher?

The cores ratings don't really have anything to do with how much offset they can run, that's different for every core on every CPU sample, you gotta test. My 5900X did -2 and -5 on its best cores, yet the next best core took -30 and clocked higher than the preferred cores. My old 5700G did -10 across the board, my new one does -15. My 5800X3D does -25 all cores.

 

[AusWolf]

Updated my BIOS from version F6B to F7A. After the update, my previously stable memory OC settings were suddenly giving me all kinds of issues (unable to POST, freezing in BIOS, Windows memory_management BSOD, etc.)

Turns out this setting was the culprit:

It was set on "Auto" on the newest BIOS, which probably meant that it was Enabled by default. So I disabled it and all the issues I had were gone! It was probably causing all kinds of memory training issues when it was enabled.

That setting has given me trouble with my MSi board too.

Is this a new AGESA feature to address long AM5 POST times? Seems like a good solution......if it worked...lol

It's in my board's BIOS since launch.

 

[Chomiq]

Done any proper testing with corecycler, OCCT or ycruncher?

The cores ratings don't really have anything to do with how much offset they can run, that's different for every core on every CPU sample, you gotta test. My 5900X did -2 and -5 on its best cores, yet the next best core took -30 and clocked higher than the preferred cores. My old 5700G did -10 across the board, my new one does -15. My 5800X3D does -25 all cores.

Well duh, obviously I haven't

Thanks for the tip, I'll give it a go. Already got myself corecycler, boosttester and PBO2 tuner.

 

[AusWolf]

Now I started having random system restarts with no definite trace in the event viewer. I'm getting the feeling that either my motherboard, or my RAM is faulty. Or could it be the CPU?

Why did I decide to give MSi another chance after being burned so many times?!

 

[Mussels]

Now I started having random system restarts with no definite trace in the event viewer. I'm getting the feeling that either my motherboard, or my RAM is faulty. Or could it be the CPU?

Why did I decide to give MSi another chance after being burned so many times?!

Tried lowering the mem clock/IF to see if it helps?

XMP on, but then lower it to something between XMP and JEDEC

 

[tabascosauz]

Now I started having random system restarts with no definite trace in the event viewer. I'm getting the feeling that either my motherboard, or my RAM is faulty. Or could it be the CPU?

Why did I decide to give MSi another chance after being burned so many times?!

Is your RAM actually stable in TM5? At least 3 cycles 1usmusv3 to get a rough idea. Shouldn't take u more than an hour

 

[Tech_fanatic]

Updated my BIOS from version F6B to F7A. After the update, my previously stable memory OC settings were suddenly giving me all kinds of issues (unable to POST, freezing in BIOS, Windows memory_management BSOD, etc.)

Turns out this setting was the culprit:

It was set on "Auto" on the newest BIOS, which probably meant that it was Enabled by default. So I disabled it and all the issues I had were gone! It was probably causing all kinds of memory training issues when it was enabled.

I have also installed the latest BIOS from the OEM. The motherboard takes 30sec plus just for the memory training and for the boot screen to appear on any memory setting other than the stock JEDEC 4800. This happens even with the XMP3 profile of 6000 on every boot/reboot. So I am forced to stay with 4800 atm. Also there is no such setting on my motherboard for memory as shown in your screenshot. Can't wait 30 secs just for the monitor to turn ON. Come on Ryzen community let's get this to AMD asap

 

[MagnyCours]

Mine is nested inside the AMD CBS menu. I'll post some pictures below, hope it'll help find yours.

 

[tabascosauz]

Mine is nested inside the AMD CBS menu. I'll post some pictures below, hope it'll help find yours.

This one of those times I wish GB and Asrock would get with the times and add a search function to their BIOS, like Asus and MSI have done for years