Discussion Zen 5 Builders thread

Page 35 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

DaaQ

Golden Member
Dec 8, 2018
1,772
1,252
136
Some sort of bug in the software, showed 223 C or -221 F. Colder than LN2 lol.
 

Seba

Golden Member
Sep 17, 2000
1,596
257
126
Some sort of bug in the software, showed 223 C or -221 F. Colder than LN2 lol.
Did you made the conversion or those are the numbers shown (possibly when switching the displayed value between C and F)?

223 C would be 433.4 F.

That's very hot, not very cold.
 
Last edited:

Det0x

Golden Member
Sep 11, 2014
1,417
4,803
136
Have received my new G.SKILL 2x16GB 6000 CL26 EXPO memory set

First thing i did was to check how low i could go with VDD voltage while still running stock EXPO primary timings, to check how much "headroom" i had against stock settings..
Kinda seems like the sticks ive gotten of this binned SKU is especially superbinned 🤣

Only require 1.34 VDD for a quick karhu run @ 6000 CL26
 

Det0x

Golden Member
Sep 11, 2014
1,417
4,803
136
Have completed my initial minimum VDD requirement for 6400MT/s CL26 testing on about half of my memory, so will share my preliminary results as promisted

The following memory sticks have put to the test:

Trident Z5 Neo RGB DDR5-6000 CL26-36-36-96 1.40V @ custom heatsinks (new generation PCB*)
Trident Z5 Royal Neo DDR5-8000 CL38-48-48-128 1.45V @ custom heatsinks
Trident Z5 Neo RGB DDR5-8000 CL38-48-48-128 1.45V
Lexar ares gen2 6000 CL26 @ 1.45v (new generation PCB)

My interest was mainly to discover how tightly these sets were binned, and if there was a large difference between the SKUs
I picked the 6400MT/s memory speed at tightish timings since i view this a middle of the road approach which all A-die should be able to run, and other overclockers also can use this as a comparison for their sticks if they wish to. Please note tCL is the main timing that can be forced with VDD, and will also be the limiting factor when voltage limited/constrained. Stress-test of choice was karhu which i started when i went to bed for the night or work in the morning, so target runtime was around 6-9 hours.

Should also be said that i'm fully aware this is only a sample size of 1 per SKU, so it dont mean too much, but i have to work with what i have

Half of the memory have been run on the ASUS APEX while for the other half on my livingroom media-computer running a ASUS GENE+2233FCK 9800X3D sample .
But at this lowish memoryspeed motherboard should not really matter, as long as the cpu is is capable and the motherboard is not total junk.
(and 3 different cpus were used for this, but the only metric i'm looking at here is VDD+stability at given mem settings)

That's enough talking, over to some data

My new G.SKILL 6000 CL26 seems to be out of this world and managed to run these settings at only 1.47 VDD which i have to say i find really impressiv, these new sticks really are the next generation.


Next up we have the Royal Neo 8000 CL38 which used to be my "primary memory" before. It needed 1.590 VDD to keep stability


My regular Trident 8000 CL38 fared the worst of the bunch and needed 1.64 VDD


And lastly we have the Lexar 6000 CL26 which also did pretty good, coming in at 1.56 VDD.
One strange thing about this printer memory is that it absolutely hate to run 2:1 mode on the GENE/ASUS, even at very slow timings and 8000MT/s speeds, which all the other memory kits above ran with ease
. System was acting totally loco and was very unstable with the lexar memory installed running 2:1 mode while in 1:1 mode it was totally fine.
When Chew was active he found kinda the same, Lexar was a good match with a Gigabyte Tachyon motherboard and a bad match for ASUS GENE motherboard. No idea what's the cause of this, only rreporting my experience ..
So can only recommed this memory if you plan to run 1:1 mode on ASUS motherboard, good thing is that is pretty cheap compared to the rest.


Following memory kits waiting to get added to the comparison:

2 different sets of G.SKILL 6400 CL30
1 set G.SKILL 6000 CL26 dual rank (working on it as we speak)
1 old set of G.SKILL 6400 CL32 dual rank

And i might throw in some kleev 7600, teamgroup UD5 7800 and T-FORCE XTREEM 8000 depending on the interest (y)

* = the new 6000 CL26 PCB (Credit goes to invictus for taking nice photos, first thing i did was drown mine in painthinner and installing the new heatspreaders, without taking any pictures)
 

Det0x

Golden Member
Sep 11, 2014
1,417
4,803
136
And lastly 8000MT/s with tight timings

This new G.SKILL 6000 CL26 set runs 8000MT/S CL30 at the same voltage my old memory-kit needed for CL32 pretty much..

For 8000MT/s CL30 only 1.71 VDD was required, which i would say is fully daily'able, especially with the custom heatsinks only reaching ~38 degrees on the memory
367.7mb/s in karhu for this memoryspeed i believe is a new "highscore" for me

Settings is not fully minimized/maxed out either.. tRCDWR, tRC and tRFC can probably go lower, but i dont want to spend too much time working on the 8000MT/s strap.
Have moved on to 8300-8400MT/s and working on a profile there now.
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,983
3,432
126
Guys Confirmed ThreadRippers all have internal fuses now which trigger even on activation of EXPO.

There is still a confusion on what part of your warrenty remains tho, as AMD says it doesn't invalidate all warrenty, only those associated with overclocking the CPU itself, but you will blow that fuse even if you activate only EXPO.



This is a sad sad moment for us overclockers.
I feel you guys on normal Ryzen's or the Ryzen 9 will get these fuses as well soon.
 

Hulk

Diamond Member
Oct 9, 1999
5,061
3,545
136
9950X3D on March 12. $699 on MC. I hope I can resist. I don't love paying full MSRP but it could be fun, depends on if the 3D cache helps any of the apps I use.

I would NEVER pay a "scalpers" fee but paying full "real" price is supporting AMD and the vendor, I'm okay with that.

Don't buy from scalpers. Let them buy up stock and sit on it until availability is there at MSRP or below and they get stuck with their purchases they can't not only sell for a profit but they actually lose money once fees and taxes are factored in.

This is what I wish Intel, nVidia, and AMD would do with the next big release. First, on the actual launch day put out limited availability and MSRP just a bit higher than actual so the scalpers buy them all up.

Then like 2-3 days later FLOOD the market with the parts at slightly lower prices. Scalpers are stuck with a bunch of parts that they paid higher than the going price.

It's a win-win for the manufacturers. They sell a bunch of parts at high prices AND stick it to the scalpers. Then reward their loyal customers with lower prices and plenty of supply.

It could be a good trick. Do it on one release but not the next. Keep the scalpers guessing. Play the game.

Hey, what the scalpers are doing is legal so the manufacturers can strike back using similar legal tactics.
 
Jul 27, 2020
23,610
16,577
146
9950X3D on March 12. $699 on MC. I hope I can resist. I don't love paying full MSRP but it could be fun, depends on if the 3D cache helps any of the apps I use.
I wish you used some open source app with a built-in benchmark for your workflow. I could easily test it for you on my Epyc with 256MB L3 cache.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
26,824
15,826
136
9950X3D on March 12. $699 on MC. I hope I can resist. I don't love paying full MSRP but it could be fun, depends on if the 3D cache helps any of the apps I use.

I would NEVER pay a "scalpers" fee but paying full "real" price is supporting AMD and the vendor, I'm okay with that.

Don't buy from scalpers. Let them buy up stock and sit on it until availability is there at MSRP or below and they get stuck with their purchases they can't not only sell for a profit but they actually lose money once fees and taxes are factored in.

This is what I wish Intel, nVidia, and AMD would do with the next big release. First, on the actual launch day put out limited availability and MSRP just a bit higher than actual so the scalpers buy them all up.

Then like 2-3 days later FLOOD the market with the parts at slightly lower prices. Scalpers are stuck with a bunch of parts that they paid higher than the going price.

It's a win-win for the manufacturers. They sell a bunch of parts at high prices AND stick it to the scalpers. Then reward their loyal customers with lower prices and plenty of supply.

It could be a good trick. Do it on one release but not the next. Keep the scalpers guessing. Play the game.

Hey, what the scalpers are doing is legal so the manufacturers can strike back using similar legal tactics.
You do realize this hap been OOS at newegg and Amazon since it was released, so sure, PAY IT !!!
 

CP5670

Diamond Member
Jun 24, 2004
5,652
757
126
Those CL26 packs look impressive. I wonder how much of a difference it makes with game fps though compared to CL30.
 
Reactions: lightmanek

IEC

Elite Member
Super Moderator
Jun 10, 2004
14,566
5,962
136
9950X3D on March 12. $699 on MC. I hope I can resist. I don't love paying full MSRP but it could be fun, depends on if the 3D cache helps any of the apps I use.

I would NEVER pay a "scalpers" fee but paying full "real" price is supporting AMD and the vendor, I'm okay with that.

Don't buy from scalpers. Let them buy up stock and sit on it until availability is there at MSRP or below and they get stuck with their purchases they can't not only sell for a profit but they actually lose money once fees and taxes are factored in.

This is what I wish Intel, nVidia, and AMD would do with the next big release. First, on the actual launch day put out limited availability and MSRP just a bit higher than actual so the scalpers buy them all up.

Then like 2-3 days later FLOOD the market with the parts at slightly lower prices. Scalpers are stuck with a bunch of parts that they paid higher than the going price.

It's a win-win for the manufacturers. They sell a bunch of parts at high prices AND stick it to the scalpers. Then reward their loyal customers with lower prices and plenty of supply.

It could be a good trick. Do it on one release but not the next. Keep the scalpers guessing. Play the game.

Hey, what the scalpers are doing is legal so the manufacturers can strike back using similar legal tactics.

I agree. I have never paid a scalper and refuse to out of principle.

I expect CPU availability will be better than gaming GPU availability. 2x 70.6mm^2 dies + 3D vCache is still quite a lot less of a wafer than 350mm^2+ for a single consumer GPU and even with 3D would expect yields and profit is better vs a single RX 9070XT.
 

MS_AT

Senior member
Jul 15, 2024
556
1,170
96
I could easily test it for you on my Epyc with 256MB L3 cache.
I am afraid your 256MB is not on one CCD and this makes a difference

How about some runs at synced golden rule 8800MT/s 2200mhz FCLK instead
Have you tried running llama.bench? While its complicated by the fact you would need to pick some model that fits in memory (maybe llama 3.2 8B params, q6_k_m), the token geneeration score should be directly proportional to the read memory BW. In case you would be interested llama.cpp github should have some documentation on how to use llama.bench and the binaries itself https://github.com/ggml-org/llama.cpp
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,983
3,432
126
So my brother is like.. can you build me something better then RGB...

Oh boy... you guys are gonna roll on the floor when you see this. I actually 1 up'd RGB... Hyte Y70 Touch Infinite, and the Thermalaright warframe with LCD's. Yep... pictures come tomorrow.

Excuse me while i dig my own grave...

But this is the completed 9800X3D.

And like promised the pictures.



Need to do some wire management, but im waiting for his videocard.
 
Last edited:

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,983
3,432
126
It would be a killer build with killer female assassins instead of goldfish.

Maybe the Kill Bill Build?

im actually gonna troll him and put a Loli Kami gif later on.
The thing takes mp4, as backgrounds as well, so you can have that.

But the case is honestly fun... it really came down to either this or the snowblind case.
But i think the Hyte is more reasonable vs the snowblind.

This is snowblind btw:

And here is the actual case:
 
Reactions: igor_kavinski
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |