Question Zen 4 builders thread

Page 117 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Hotrod2go

Senior member
Nov 17, 2021
349
232
86
I don't think the IHS is really a big an issue as we think it is, most in depth testing showed thinning it out only netted a few degrees at best. I am using $20 coolers that clip on the two sides, got tired of the tear it all down mounting nonsense with "fancy" coolers and the cooling limit has to do with how high the die temp is allowed to get. If you set it at 85 ish or under, it will greatly help that aspect; running up to 90+ is not ideal.

My 7600X is in mothballs for a bit, but I set it like the 7800X3D; ~75% limits for PBO settings, -15 all core curve offset, SoC voltage at 1.055 and power usage and temps go way down after that.

Here is the 7800X3D, idle ish, RDP to the PC, so it is a little off on "best" looking temps, on a $20 cooler. Core temp limits were still reached in gaming on a 280mm AiO, so, I just went back to air :|

View attachment 106056
95 C is the acceptable maximum temp for Zen 4 according to AMD, just as it is with Zen 5, so with the extra think IHS they have engineered AM5 line of chips to withstand that for daily usage. AM6 will change that I think, but that's yrs off. However like you, I don't like that either but it is what it is. Your low SOC voltage is because your running only 2029 FLCK (29 MHz over is probably due to minor variations from software monitoring) & approx 5.05GHz maximum for 7800X3D is typical as they do not overclock well even with PBO tuning unlike the non X3D chips, but I'm sure you know that anyway. Do you have the BCLK locked at 100MHz in bios?
The usage model for this 7600X system I have will not benefit from PBO tuning as its just for daily desktop non gaming use, that's why I don't have a dGPU in it.
If you look on the market, there are tools to delid these chips & run direct to die, so for them to be manufactured in the first place there must be enough demand out there for them to bother with that aspect in the first place. Take a look at the Themal Grizzly line of products for this as one example. In fact they also make them for Intel's socket 1700 - overall though its telling us that both AMD (via TSMC) & Intel, despite advancements in shrinkage of transistors, still come up against the laws of physics.
 
Last edited:
Reactions: lightmanek

B-Riz

Golden Member
Feb 15, 2011
1,591
754
136
95 C is the acceptable maximum temp for Zen 4 according to AMD, just as it is with Zen 5, so with the extra think IHS they have engineered AM5 line of chips to withstand that for daily usage. AM6 will change that I think, but that's yrs off. However like you, I don't like that either but it is what it is. Your low SOC voltage is because your running only 2029 FLCK (29 MHz over is probably due to minor variations from software monitoring) & approx 5.05GHz maximum for 7800X3D is typical as they do not overclock well even with PBO tuning unlike the non X3D chips, but I'm sure you know that anyway. Do you have the BCLK locked at 100MHz in bios?
The usage model for this 7600X system I have will not benefit from PBO tuning as its just for daily desktop non gaming use, that's why I don't have a dGPU in it.
If you look on the market, there are tools to delid these chips & run direct to die, so for them to be manufactured in the first place there must be enough demand out there for them to bother with that aspect in the first place. Take a look at the Themal Grizzly line of products for this as one example. In fact they also make them for Intel's socket 1700 - overall though its telling us that both AMD (via TSMC) & Intel, despite advancements in shrinkage of transistors, still come up against the laws of physics.

Ahhh, yes and no?

95*C maybe fine technically, but some of us are old school and it gives us the nervous sweats "we can tweak it down, right?" 🤣

The thicker IHS is to allow the end users to keep using AM4 coolers, waste not want not?

Someone actually did the heat transfer math on the IHS of Zen4 vs Zen5 and found that the thicker IHS of Zen5 was a nothing burger, higher core temps are just due to the nature of the cpu silicon transistor size and density now. (No, I cannot remember where I saw it right now 😑) A shaved down Zen5 IHS would not magically make it run cooler with the same cooling setup.

In general, we are fighting physics with CPU's of the last 5+ years, having to cool a very tiny and very hot area using inexpensive mass produced parts.

Delidding is interesting, but really more a novelty.

I and many others lived the bad old days before IHS's, I fried at least one Socket A Athlon XP with a bad cooler mounting. (´。_。`)

I run 2033 per BuildZoid lazy DDR4 6000 OC settings, and SoC voltage is about the lowest I can go and be stable per prime95 stress testing; at least for that 7800X3D chip, I never read anything that indicated BCLK really needs to be locked...

A 7600X with power tuned settings would totally be worth it for even daily non intensive use, lower power draw, lower temps, etc.

The great thing about these chips is they can be tweaked to use less power and not lose substantial MT performance.

A reddit user did an in depth test of their 7800X3D, limiting max die temp is one of the easiest ways to "tune" the chip and not have to use a big AIO or custom loop for the "max perf low temps".

I tried a pure max temp limit, but got better performance and temps by using the ~75% OOB PBO limits, -15 all core offset and setting SoC to 1.055; that was stress tested with prime95 mixed mode for 24 hours and passed. I had tried a -20 offset after that, but that was too low and had some stability issues.


Conclusions​


The 7800X3D is great out of the box. Tweaking PBO+CO to limit thermals/power consumption— while increasing performance over stock—is a game within itself (min/max, anyone?). Just don’t get caught up trying to chase online benchmarks. The real-world difference is negligible.


That being said, the observed difference between the lowest Cinebench score to the highest (stable) was a mere 2.8%. That’s not much potential you are leaving on the table should you decide to go with all-core, per-core, fixed wattage, and/or applying a thermal limit.


I didn't test in a temperature-controlled room or utilize state-of-the-art equipment, but the observed results confirm my hypothesis: the 7800X3D running at/above 85°C is by design. Over 80°C and you add additional heat for limited performance gain. It’s up to each user to decide if the trade-off is worth it.


The last time I messed around with overclocking was an Intel Q6600 which ran at 2.4 GHz stock. I pushed it to 3.2 GHz on air cooling, but settled for 3.0 GHz daily usage. That’s a 33% and 25% bump, respectively. Nowadays, we're pumping less juice for more power. My have the times changed!


Technology has advanced tremendously since, and that includes online resources available to PC enthusiasts. Hopefully this post summarizes all the information that took me days to research and compile. Lastly, we should all be glad that a chip like the 7800X3D exists and makes high-performance gaming accessible to all who can afford it.
 

Hotrod2go

Senior member
Nov 17, 2021
349
232
86
Ahhh, yes and no?

95*C maybe fine technically, but some of us are old school and it gives us the nervous sweats "we can tweak it down, right?" 🤣

The thicker IHS is to allow the end users to keep using AM4 coolers, waste not want not?

Someone actually did the heat transfer math on the IHS of Zen4 vs Zen5 and found that the thicker IHS of Zen5 was a nothing burger, higher core temps are just due to the nature of the cpu silicon transistor size and density now. (No, I cannot remember where I saw it right now 😑) A shaved down Zen5 IHS would not magically make it run cooler with the same cooling setup.

In general, we are fighting physics with CPU's of the last 5+ years, having to cool a very tiny and very hot area using inexpensive mass produced parts.

Delidding is interesting, but really more a novelty.

I and many others lived the bad old days before IHS's, I fried at least one Socket A Athlon XP with a bad cooler mounting. (´。_。`)

I run 2033 per BuildZoid lazy DDR4 6000 OC settings, and SoC voltage is about the lowest I can go and be stable per prime95 stress testing; at least for that 7800X3D chip, I never read anything that indicated BCLK really needs to be locked...

A 7600X with power tuned settings would totally be worth it for even daily non intensive use, lower power draw, lower temps, etc.

The great thing about these chips is they can be tweaked to use less power and not lose substantial MT performance.

A reddit user did an in depth test of their 7800X3D, limiting max die temp is one of the easiest ways to "tune" the chip and not have to use a big AIO or custom loop for the "max perf low temps".

I tried a pure max temp limit, but got better performance and temps by using the ~75% OOB PBO limits, -15 all core offset and setting SoC to 1.055; that was stress tested with prime95 mixed mode for 24 hours and passed. I had tried a -20 offset after that, but that was too low and had some stability issues.

Yes, no arguments there. Photonic semiconductors will be the way of the future, fast moving electrons cause too much heat with current designs no matter how small the transistors get & they cram into approx the same die area. I also heard today that China is now leading the USA with production of this technology but take that as hearsay until you do some research.

Your right in the fact that the thick IHS, at least in Zen 5 maybe more of a nothing burger atm, at least with some FCLK overclocking I've been experimenting with today. If you go over to this thread, I found evidence of this claim for my particular chip.
 

Hotrod2go

Senior member
Nov 17, 2021
349
232
86
Today I finally took delivery of my Ryzen 5 8600G & promptly went to work installing & setting it up. After double checking the support list for my Gigabyte board & 8000 series memory support (which they do for my Corsair Vengeance RGB kit ) I was wondering why the system would not boot first time.
Turns out the IMC for 8000 series is fussy at least with Samsung B die chips on my DIMMs, despite clear unambiguous listing of the supported memory modules. I had the F21 bios installed (AGESA 1.1.0.1) but no boot with both sticks, pulled out the B2 slot stick... still no boot... pulled out the A2 stick & had one in the B2 slot only... now it boots!
I tried the USB bios flashback system for this board, but that would not work either... press the button it starts then couple seconds later dies...

Anyway, long story short, updated to F31 bios (AGESA 1.2.0.0a) now it can boot with both sticks after loading system defaults however I'm going to stay on 4800MT/S speed for now despite the RAM having 5600MT/S (XMP) profile in SPD as I'm sus the system is stable without clearing CMOS because it fails to reboot unless a cold boot process is initiated.

So this is a warning to those venturing into 8000 series APUs with Gigabyte B650 boards & Samsung B die RAM! There is nothing wrong with this RAM kit as it has been thoroughly tested with several apps under a Ryzen 7 7700 chip & found to be very stable indeed with XMP profile.
 

Hotrod2go

Senior member
Nov 17, 2021
349
232
86
Ok, got this problem figured with the Ryzen 5 8600G system in my signature & the Corsair Vengeance RGB 5600MT/S kit. Seems the board was undervolting the XMP profile VDD & VDDQ voltage lines. Vdroop is probably the more appropriate term here. But anyway its sorted now.
Ran PCMark10 default run with cpu on defaults & XMP profile loaded with correction to the RAM voltage lines. 1.26v for VDD & VDDQ seems to fix it instead of 1.25v default for both. This was the only way to get the system to reboot successfully with this RAM kit @ XMP profile timings.


Some perspective on the results in comparison with the rest of the scoreboard;



What's fascinating about this chip is the way the FCLK & UCLK speeds fluctuate depending on system usage at the time. Didn't see that in the 7000 or 9000 series chips I have.
Considering the system is more productivity focused, I'm pleased with this so far, but the tweaker in me can't help himself so more OC results will be incoming soon...
 
Last edited:

Hotrod2go

Senior member
Nov 17, 2021
349
232
86
Maybe that's required to get more performance out of the iGPU? What do the bandwidth figures look like for your three CPUs
Can do that, but will have to be at 5200MT/S CL40 for all systems as the Adata B die kit in the 7600X rig is not a good over clocker ( new kit coming next week) to be stable pass that speed & timings without too much VDD. Just need a bit of time to get all 3 results.

Edit: not going to happen atm.
 
Last edited:

biostud

Lifer
Feb 27, 2003
18,606
5,301
136

Abwx

Lifer
Apr 2, 2011
11,514
4,301
136
When people aren't buying the 7900X3D, but you don't really want to sell the 7600X3D, but also have a pile of 6-core 3XD chiplets sitting on the shelves collecting dust.

By the time they know the demand for each SKU, so they are not manufacturing 7900X3D in random quantities, and given its unmatched price/perf ratio this CPU sell better than one would think, it s just that Micro Center has several bns revenue and it s worthy to provide them such an exclusivity.

 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
29,444
24,111
146
Raptor Lake, more like Leper Lake. MF sold as many ancient 3200G as they did 14900K.

Zen 4 3D accounted for 1130 sales. 3D all together = 1370.

A point to consider when looking at retail sales, that has frequently happened the last year; If it isn't top 5, sales are often only a tiny fraction of the top seller. The 7800X3D for that week was 930 while the 10 spot is 60. That's an enormous gap. Amazon U.S. the top 10 are all AMD with Zen 4 being 5 of those. That's a nice improvement from last year. No one cares about Zen 5 at current prices.

I was initially surprised the 7800X3D is so wildly popular. Then it occured to me that in adjusted dollars its MSRP was almost the same as the legendary i7-2600K was on release. At the present price it is Devil's Canyon i5 K territory. Given that context it is not surprising at all.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
26,049
15,191
136
Raptor Lake, more like Leper Lake. MF sold as many ancient 3200G as they did 14900K.

Zen 4 3D accounted for 1130 sales. 3D all together = 1370.

A point to consider when looking at retail sales, that has frequently happened the last year; If it isn't top 5, sales are often only a tiny fraction of the top seller. The 7800X3D for that week was 930 while the 10 spot is 60. That's an enormous gap. Amazon U.S. the top 10 are all AMD with Zen 4 being 5 of those. That's a nice improvement from last year. No one cares about Zen 5 at current prices.

I was initially surprised the 7800X3D is so wildly popular. Then it occured to me that in adjusted dollars its MSRP was almost the same as the legendary i7-2600K was on release. At the present price it is Devil's Canyon i5 K territory. Given that context it is not surprising at all.
Zen 5 is probably only popular with those that actually regularly use avx-512 (the DC people). But Intel in general is in a dead end. High end degrade and die, low end are dead end, so AMD rules right now in desktop and server.
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
29,444
24,111
146
Zen 5 is probably only popular with those that actually regularly use avx-512 (the DC people). But Intel in general is in a dead end. High end degrade and die, low end are dead end, do AMD rules right now in desktop and server.
Yeah, I saw you DC folks are doing alright with the new shiny. Retail is driven by gaming, which is why the 3Ds sell so well. The 7800X3D may end up the best selling retail CPU ever unless the 9800X3D is something special. Leper Lake is dead in retail.
 

biostud

Lifer
Feb 27, 2003
18,606
5,301
136
Raptor Lake, more like Leper Lake. MF sold as many ancient 3200G as they did 14900K.

Zen 4 3D accounted for 1130 sales. 3D all together = 1370.

A point to consider when looking at retail sales, that has frequently happened the last year; If it isn't top 5, sales are often only a tiny fraction of the top seller. The 7800X3D for that week was 930 while the 10 spot is 60. That's an enormous gap. Amazon U.S. the top 10 are all AMD with Zen 4 being 5 of those. That's a nice improvement from last year. No one cares about Zen 5 at current prices.

I was initially surprised the 7800X3D is so wildly popular. Then it occured to me that in adjusted dollars its MSRP was almost the same as the legendary i7-2600K was on release. At the present price it is Devil's Canyon i5 K territory. Given that context it is not surprising at all.
Yeah, since most desktop users who buy custom parts are probably gamers, they are really not that interested in paying extra for the MT performance of the 14900K, the added heat, and now degradation. Pricewise the 14700K and the 7800X3D is the same here in Denmark (with the 7800X3D increased slightly in price the last month and the 14700K declined, for some reason ) So why choose the 14700K when you are a gamer? On the other hand the regular 7700 is 40% cheaper than the 7800X3D, so you could get a pretty decent budget gaming setup going with that.
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
29,444
24,111
146
Yeah, since most desktop users who buy custom parts are probably gamers, they are really not that interested in paying extra for the MT performance of the 14900K, the added heat, and now degradation. Pricewise the 14700K and the 7800X3D is the same here in Denmark (with the 7800X3D increased slightly in price the last month and the 14700K declined, for some reason ) So why choose the 14700K when you are a gamer? On the other hand the regular 7700 is 40% cheaper than the 7800X3D, so you could get a pretty decent budget gaming setup going with that.
No probably about it, gaming drives retail CPU sales. When we say RIP Leper Lake, it stands for rust in peace. I know that isn't what happens to them, but it's funnier that way. Just to interdict anyone pushing their glasses up and preparing to "ACKCHYUALLY..." me.
 

Ranulf

Platinum Member
Jul 18, 2001
2,509
1,570
136
LMAO 7600X3D is another MC exclusive

https://www.pcmag.com/news/meet-the-ryzen-5-7600x3d-amd-teams-up-with-micro-center-on-3d-v-cache-cpu

edit: I am laughing because there will be many salty about it.

Just saw this news via a video by gamermeld from yesterday. Apparently HUB did a test with a simulated 6core 3dcache by turning off the 6 non 3d cache cores in a 7900X3d but I can't find the video in a quick search. The big comparison there being a 7600x3d theoretically beats out a 14700k in price/perf. Neat or something. It is Intel rusting chips, who cares.

The theoretical performance numbers look nice but at $299 it really only makes sense if you can't get a good deal on the 7800x3d which has gone up in price and is low stock in the US, online at least with prices hitting $400 or so. Why pay $299 for 3d cache when a 7600X is $194? Let alone the 7500F outside of the US. So just another MC sales gimmick to get rid of some excess chips that didn't work out as a 12core etc..
 
Reactions: Tlh97
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/    |