• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Are you using an AMD Ryzen X3D CPU with 3D V-Cache?

Are you using an AMD Ryzen X3D CPU with 3D V-Cache?

  • Yes, Zen 3 (5000X3D)

    Votes: 4,330 13.6%
  • Yes, Zen 4 (7000X3D)

    Votes: 3,038 9.5%
  • Running a classic Ryzen

    Votes: 14,775 46.2%
  • Running Intel

    Votes: 9,807 30.7%

  • Total voters
    31,950
  • Poll closed .
Joined
Jun 6, 2022
Messages
622 (0.85/day)
I understood very well, but you don't get the ideas.
For overclocking, chipset B motherboards are not good because they primarily block CPU overclocking. They are an option only if you are not interested in overclocking, but in system stability.
What you didn't understand and still don't is that DDR5 only works on the Gear 2. It's one of the reasons why a processor with DDR4 RAM memories clocked at 3600MHz achieves the same performance as the same processor with DDR5 RAM memories clocked at 5600MHz. We are talking strictly about gaming and with very powerful video cards.
In short, where you see DDR5 6000 (obtained only with Gear 2 because Gear 1 far exceeds the limits of the controller) the frequency of the memory controller is the same as when using DDR4 at 3000 MHz.

My locked processor allows me 3600MHz Gear 1, equivalent to 7200MHz Gear 2.
Where you see DDR4 4400MHz, these frequencies are possible by pushing the voltages to very dangerous values and the stability also suffers. They are for exposure on forums, not for intense activity and online games where a reset destroys your work.

As an idea, the voltage of the memory controller and the RAM memories are also unlocked on chipset B motherboards. So is the frequency of these RAM modules. I have 12500 and 13500 and both work perfectly with the memories clocked at 3600MHz Gear 1. I only tested the 12500 for a few days, but the 13500 runs stable as a rock since February. This is the memory limit, I don't know if processors can do more, but I'm not interested in a 0.01% increase in performance and 1000% system instability.
On both processors, VDDQ TX Voltage (controller voltage) is set at 1.1V. The nominal voltage is 1.2V if XMP is not activated and 1.35V if it is activated.
 
Last edited:
Joined
Nov 29, 2007
Messages
979 (0.16/day)
Location
Netherlands
What you didn't understand and still don't is that DDR5 only works on the Gear 2.
On AMD platforms it does run in 1:1 or 1:2 for sure and has been tested which is equal to gear 1 and 2 on intel so that is just blatant misinfo you could have googled quickly. No clue how it is on intel but has nothing to do with DDR5 not being able to. https://www.igorslab.de/en/ryzen-70...ctice-test-with-benchmarks-recommendations/5/

On topic still running my 3600, might consider upgrading to a 5800X3D soon grabbing it cheap from the used market. Already seen them go for around 200-250 euros used which is a steal. Maybe even wait it out till other gens not really in a rush to upgrade the lifespans on cpu's these days are great.
 
Joined
Jun 6, 2022
Messages
622 (0.85/day)
On AMD platforms it does run in 1:1 or 1:2 for sure and has been tested which is equal to gear 1 and 2 on intel so that is just blatant misinfo you could have googled quickly. No clue how it is on intel but has nothing to do with DDR5 not being able to. https://www.igorslab.de/en/ryzen-70...ctice-test-with-benchmarks-recommendations/5/
That's right, you really have no idea about Intel. In the link you will notice the AMD Fabric Clock, which is neither Gear 1 (the frequencies are lower) nor Gear 2 (the frequencies are higher).
For AMD's Fabric Clock to be identical to Gear 1, the frequency had to be 3000 MHz for DDR5 6000MHz. I see 2000 MHz. :cool:

Even the moderator doesn't really know much about Intel because here he says that the differences are due to Gear 2 versus Gear 1. The differences are because two different processors are compared: i5-13400 versus i9-13900K. As even with Intel you cannot reach frequencies of 3000MHz with the memory controller, all tests with DDR5 are done only with Gear 2 (for 6000MHz, the memory controller reaches 1500MHz). Only for DDR4 you can use Gear 1. The difference between Gear 1 and 2 is not the writing/reading/copying speed, here the impact is minimal. The big impact is on latency.

----------------------------------------
Moderator's statement (link):
ffff.jpg

--------------------------------------------

And the comparison between Gear 1 and 2 on my system. It is clear where the losers are.
gear 1.jpg
gear 2.jpg


In conclusion, please avoid contradicting someone on subjects that you do not master.
 
Last edited:

deb

Joined
Jul 5, 2023
Messages
24 (0.07/day)
Processor Ryzen 5 5500
Motherboard ASRock X570 Steel Legend WiFi
Cooling Stock
Memory DDR4-3000 CL16
Video Card(s) GTX 970
Storage TEAMGROUP MP33
Display(s) Acer R240HY, Pixio PX248
Case Thermaltak P3 TG Pro Snow
Power Supply Super Flower Leadex III ARGB
Mouse Razer Deathadder V2 Pro Wireless
Keyboard K38
That's right, you really have no idea about Intel. In the link you will notice the AMD Fabric Clock, which is neither Gear 1 (the frequencies are lower) nor Gear 2 (the frequencies are higher).
For AMD's Fabric Clock to be identical to Gear 1, the frequency had to be 3000 MHz for DDR5 6000MHz. I see 2000 MHz. :cool:

Even the moderator doesn't really know much about Intel because here he says that the differences are due to Gear 2 versus Gear 1. The differences are because two different processors are compared: i5-13400 versus i9-13900K. As even with Intel you cannot reach frequencies of 3000MHz with the memory controller, all tests with DDR5 are done only with Gear 2 (for 6000MHz, the memory controller reaches 1500MHz). Only for DDR4 you can use Gear 1. The difference between Gear 1 and 2 is not the writing/reading/copying speed, here the impact is minimal. The big impact is on latency.

----------------------------------------
Moderator's statement (link):
View attachment 314385
--------------------------------------------

And the comparison between Gear 1 and 2 on my system. It is clear where the losers are.
View attachment 314386View attachment 314387

In conclusion, please avoid contradicting someone on subjects that you do not master.
You rock
 
Joined
Feb 24, 2021
Messages
122 (0.10/day)
System Name Upgraded CyberpowerPC Ultra 5 Elite Gaming PC
Processor AMD Ryzen 7 5800X3D
Motherboard MSI B450M Pro-VDH Plus
Cooling Thermalright Peerless Assassin 120 SE
Memory CM4X8GD3000C16K4D (OC to CL14)
Video Card(s) XFX Speedster MERC RX 7800 XT
Storage TCSunbow X3 1TB, ADATA SU630 240GB, Seagate BarraCuda ST2000DM008 2TB
Display(s) AOC Agon AG241QX 1440p 144Hz
Case Cooler Master MasterBox MB520 (CyberpowerPC variant)
Power Supply 600W Cooler Master
I bought a 5800X3D earlier today.
It arrives in 2 days.
I put myself down as using it even though I don't actually have it yet.
 
Joined
Jan 28, 2023
Messages
23 (0.05/day)
Location
Toronto, Canada
System Name Codsworth
Processor i9 - 10900F
Motherboard Z490 MSI MEG ACE
Cooling Assassin III
Memory 96gb Fury Beast DDR4 @ 3600/CL18
Video Card(s) Suprim X 3080 Ti
Storage 970 Evo Plus (2TB), P3 Plus (4TB), WD Blue M.2 (1TB), 870 QVO (4TB), Seagate HDD (8TB), WD HDD (8TB)
Display(s) Alienware AW3821DW (3840x1600 - 144hz), Asus Tuf VG27QL1A (1920x1080 - 144hz)
Case Meshify 2
Audio Device(s) Meredo 2.1Ch Soundbar (150w)
Power Supply EVGA Supernova G3 850w
Mouse Corsair Katar Pro Wireless
Keyboard ROG Strix Scope TKL (MX Red)
Still rolling on this 10900f, I regret the F part but that's about it.

Given my monitor, any top-end gpu that I could buy in the next 5 years will probably be a bottleneck at 4k-like resolutions.

A platform upgrade will probably be due when I get 4k 240hz monitor, and that'll only happen when they get common enough that multiple brands have them.
 

Keullo-e

S.T.A.R.S.
Joined
Dec 16, 2012
Messages
11,312 (2.70/day)
Location
Finland
System Name 4K-gaming
Processor AMD Ryzen 7 5800X
Motherboard Gigabyte B550M Aorus Elite
Cooling Custom loop, Corsair ML/LL fans
Memory 48GB Kingston Fury DDR4-3200
Video Card(s) Asus GeForce RTX 3080 TUF 10GB @ +150/700
Storage ~4TB SSD + 6TB HDD
Display(s) Acer XV273K 4K120 + Lenovo L32p-30 4K60
Case Corsair 4000D Airflow White
Audio Device(s) Asus TUF H3 Wireless
Power Supply EVGA Supernova G2 750W
Mouse Logitech MX518 + Asus TUF P1 mousepad
Keyboard Roccat Vulcan 121 AIMO
VR HMD Oculus Rift CV1
Software Windows 11 Pro
Benchmark Scores It runs Crysis remastered at 4K
Will be using a "classic" Ryzen in the future as well, ordered a 5800X as is was on sale (190EUR).
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.14/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
I understood very well, but you don't get the ideas.
For overclocking, chipset B motherboards are not good because they primarily block CPU overclocking. They are an option only if you are not interested in overclocking, but in system stability.
What you didn't understand and still don't is that DDR5 only works on the Gear 2. It's one of the reasons why a processor with DDR4 RAM memories clocked at 3600MHz achieves the same performance as the same processor with DDR5 RAM memories clocked at 5600MHz. We are talking strictly about gaming and with very powerful video cards.
In short, where you see DDR5 6000 (obtained only with Gear 2 because Gear 1 far exceeds the limits of the controller) the frequency of the memory controller is the same as when using DDR4 at 3000 MHz.

My locked processor allows me 3600MHz Gear 1, equivalent to 7200MHz Gear 2.
Where you see DDR4 4400MHz, these frequencies are possible by pushing the voltages to very dangerous values and the stability also suffers. They are for exposure on forums, not for intense activity and online games where a reset destroys your work.

As an idea, the voltage of the memory controller and the RAM memories are also unlocked on chipset B motherboards. So is the frequency of these RAM modules. I have 12500 and 13500 and both work perfectly with the memories clocked at 3600MHz Gear 1. I only tested the 12500 for a few days, but the 13500 runs stable as a rock since February. This is the memory limit, I don't know if processors can do more, but I'm not interested in a 0.01% increase in performance and 1000% system instability.
On both processors, VDDQ TX Voltage (controller voltage) is set at 1.1V. The nominal voltage is 1.2V if XMP is not activated and 1.35V if it is activated.
I get the feeling i understand these ideas very well.
The gears is something where marketing names get in the way of the actual reality, like how corsair label all their RAM in MHz and not MT/s, which is actually illegal in many countries including australia.


When different brands and generations of the same brand change the meaning of the gears, you have to accept that i'm going to keep these explanations simplified with a gear 1/2 summary and not waste my time with DRAM ratios unless its specifically needed for what's being discussed.

My information comes from myself and from IR_cow, TPUs previous motherboard reviewer and current memory reviewer. I get the feeling he's ran into these limits a lot more than you have, he knows these problems well and it's why he's had to drop motherboard reviewing - the time needed to get most RAM kits working has gone up exponentially with DDR5, especially since intel have locked the SA voltage down so drastically.
 
Joined
Jun 6, 2022
Messages
622 (0.85/day)
My information comes from myself and from IR_cow, TPUs previous motherboard reviewer and current memory reviewer. I get the feeling he's ran into these limits a lot more than you have, he knows these problems well and it's why he's had to drop motherboard reviewing - the time needed to get most RAM kits working has gone up exponentially with DDR5, especially since intel have locked the SA voltage down so drastically.
Can you show me a single review with DDR5 6000 MHz Gear 1? Only one! It doesn't matter if it's TPU or another reviewer.
Until then, I give you the configuration used by TPU for 13400, 13600K, 13900K and 13900KS.
I repeat that you cannot use Gear 1 with DDR5, maybe only if you underclock the memories below 4400MHz. Understand that you have made a grave mistake and you are making an even graver mistake by insisting that white is black.

i5-13400.jpg
13600K.jpg

13700K.jpg
i9-13900KS.jpg
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.14/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
You're pointing at Z790 boards with K CPUs and i'm telling you the problems are on B/H boards and non K CPUs.

You pay me for a 13th gen non-K CPU and i'll test it for you - theres a big difference between having knowledge of one combination of parts, and having more of them and talking to the experts who deal with the memory reviews here at TPU - these are also not something you should be derailing an AMD CPU thread about, because you're obsessed with intel being 'better'

1697090372096.png
 
Joined
Jun 6, 2022
Messages
622 (0.85/day)
Someone named Mussels is trying to deviate from a subject that he approached in ignorance of the cause.
Let us remind him that he claims that TPU tested the i5-13400 with DDR5 6000 Gear 2 because the non-K processors have the SA voltage blocked and that this is a tragedy. Mussels obviously does not know that the 12-13th processors, blocked or not, SA blocked or not, cannot use DDR5 memories with Gear 1.
.................................................................................
What Mussels says here (Sep 17, 2023):

"TPU has the 13400F here, but they test with DDR5 6000 on a Z board
TPU needed to run this in gear 2.
...
Comparing Gear 1 vs 2, same RAM - different CPU.
...
Same RAM, same board. The board name only changed because AIDA was updated to detect it correctly.
The memory latency is the same, verifying the actual RAM hasn't changed - but the read values for example are absolutely tanked by gear 2 - about a 7% loss. That 7% is a huge loss."
..................................................................................

Unfortunately, Mussels makes two serious errors:
1. He did not notice that both processors are tested with Gear 2 (I repeat that it is impossible to use Gear 1 at such frequencies even if you connect SA to a high voltage line)
2. The synthetic results in AIDA give different results for different processors, and the difference between 13400 and 13900K is huge.

Mussels cannot convince any Intel 12-13th owner with those pictures of motherboard boxes that he has relevant information about Intel processors, from their consumption in practice to their functionality. He is probably trying to convince some AMD owners who haven't had any contact with Intel since the Pentium II.

You pay me for a 13th gen non-K CPU and i'll test it for you
Why? I have two at home, each in their own system. And I need money because the 6-8KW (consumed monthly by the system with 12500) costs a fortune. One U.S. dollar! :mad:
I use 12500 since June 2022 (motherboard B) and 13500 (motherboard Z) since February 2023. Both use DDR4 (bought in 2020) at a maximum of 3600MHz Gear 1. Yes, for DDR4 you can use Gear 1 and 3600MHz is a respectable frequency for a non-k processor. I say maximum because 12500 switched to a "shameful" 2666MHz CL14 Gear 1 because it does not affect the applications run with it. To my shame, I didn't notice that for days I ran with 2133 CL16 RAM because I forgot to make the changes in the BIOS during an update.. The eye didn't notice the "huge" performance penalty, I detected it by mistake with the help of HWinfo or CPU-Z, it doesn't matter.

No choice, AMD or Intel, is wrong if they meet your needs. I keep repeating it because I don't understand what you are trying to prove by blaming a processor just because it belongs to the other manufacturer.

P.S. I followed the wattmeter as I wrote on the TPU forum (12500 + igp, 16GB RAM). Jumps between 24 and 35W with an approximate average of 27-28W. Will it be good?

12500 www only (youtube, forum, news, etc.). It's just a reply to the malicious statements that for an Intel you need solar panels. Statements made on this topic.
HWiNFO64 v7.34-4930 Sensor Status [167 values hidden] 10_12_2023 11_30_38 PM.jpg
 
Last edited:
Joined
Nov 21, 2018
Messages
7 (0.00/day)
Location
Lake Dallas, TX
Processor Ryzen 5 2600
Motherboard ASUS TUF X470 Plus Gaming
Cooling Cryorig H7
Memory 32GB DDR4 3000mhz
Video Card(s) EVGA GeForce GTX 1080 Ti SC Black
Case NZXT H510i
Power Supply be quiet! Straight Power 11 750w 80+ platinum
Mouse Logitech G703 HERO
Keyboard Logitech G910 Orion Spectrum
Software Windows 11 Pro
I still have my 2600 from 2018, it’s definitely fallen off since I got it. Really eyeballing the 5600X3D though once I am able to, quite a feisty chip for $230
 
Top