Jump to content

RAM overclocking confusion

DaQuackers
 Share
Go to solution Solved by Guest,

I see what happened.

 

In the screen shot shows Infinity Fabric (North Bridge according to AIDA64 benchmark) running a 1:2 divider or "uncoupled" if you'd like to put it that way. Some boards will automatically do this. Thus the loss in latency.

 

However....

 

If you loosen the timings with a 1:2 IF to MEM divider, you might be able to run 4000mhz and this will increase bandwidth and lower latency slightly. Though the latency will stay inherently higher until you've reached a higher frequency, maybe 4400mhz.

 

Generally the Micron is ok with a little extra voltage. 1.4v is ok. Past this I recommend actively cooling them. Fan right on the sticks. Cooler brings better stability. It's just as important as the timings.

 

So force the 1:1 divider, or go on an OC climb. CL20-24-24 with XMP enabled, memory training should be enabled as well. Not sure if that Gigabyte board has training algorithms though. Something that's nice about Asus's top end boards. You can really OC the memory, so the board does matter too. 

 

GLHF!

Posted

So, I got a new processor.  Ryzen 5900x.  Technically I RMA'd the first one cause it kept crashing at stock settings so I got two of them.
The new one is... underperforming.  not by a lot.... 5-10% below average as reported by several sites.  So, knowing how sensitive ryzen is to ram I tried overclocking my ram a bit.

Here's where the story gets.... weird.  With tighter ram timings, and faster speeds, my cinebench r23 score went up.  By about 4%  from 20172 to 20933
BUT
PassMark ram bench shows my latency has gotten dramatically WORSE which makes no sense..... the cas latency is the same (wouldn't drop to cas16 stable) the secondary's are slightly better, and the speed is up from 3600 to 3800
This violates everything I _thought_ I knew about ram and overclocking it.  Also this motherboard will only set an even numbered cas latency which is weird to me coming from intel.  my last platform (7820x) I was running ram at 3400cl15

Ram rated profile - 3600mhz cl18-22-22-22-42 1.35v
Overclock profile - 3800mhz cl18-21-21-21-41 1.355v

FTR the overclock passes a full 4 loops of the memtest86 tester.  It's stable as it is.

Link to comment
https://linustechtips.com/topic/1411416-ram-overclocking-confusion/
Share on other sites
Link to post
Share on other sites
Posted

The subtimings are probably off. I haven't messed with timings extensively, but they're very particular. Each corresponds to a particular RAM procedure, and somewhat unintuitively lowering one in isolation doesn't necessarily net better response, if the timing causes it to miss a pass of another related function's timing.

 

Long and short, the main benefit comes from lowering CL, of which, changing subtimings may be necessary. However, if you can't get the CL lower, you're mostly just doing a bunch of work, with little to no real benefit, and the potential for actual detriment if it's not just right. I'd set everything back to auto, bump the clockspeed back up, and then test from there.

 

If you do want to delve deeper into timings, I'd suggest doing a lot more research to make sure you understand exactly what each timing is for and how they're related to each other, and not proceed until you're sure you're making changes that make sense.

 

CPU: AMD Ryzen 9 9950X3D · Cooler: Noctua NH-D15S Chromax.black · Motherboard: Gigabyte Auros X670 Elite AX · RAM: G.Skill Flare X5 64GB (2 x 32GB) DDR5 6000MHz CL30 · Graphics Card: Zotac NVIDIA GeForce RTX 4070 Super Twin Edge OC 12GB · Boot Drive: 1TB XPG Gammix S70 Blade NVMe SSD · Game Drive: 2TB WD SN850X NVMe SSD · PSU: Seasonic Focus GX V3 1000W 80+ Gold · Case: Fractal Design North Mesh · Monitor: MSI Optix MAG342CQR 34” UWQHD 3440x1440 144Hz · Keyboard: EPOMAKER x Aula F99 Wireless Mechanical Keyboard · Mouse: Logitech G309 Lightspeed Wireless Gaming Mouse

Link to post
Share on other sites
Posted
1 hour ago, DaQuackers said:

PassMark ram bench shows my latency has gotten dramatically WORSE which makes no sense

That's because passmark is unreliable for that.

Use AIDA64 Cache & Memory benchmark.

1 hour ago, DaQuackers said:

my last platform (7820x) I was running ram at 3400cl15

Just use 3400 CL15 if you can't get 3600MHz CL16...

 

Also what kit do you use?

I managed to overclock (on a Zen+ CPU!) a Micron E die kit from 3200MHz 16-18-18-36 to 3733MHz 14-20-14-36:

1193097160_cachemem-3733CL14.thumb.png.c3872221fedec8c439955ef669837ffe.png

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE RTX 3080 GAMING OC | 4x 8GB Micron Rev.E (D9VPP) 3800MHz 16-19-14-21-58
Link to post
Share on other sites
Posted
  • Author
2 hours ago, Vishera said:

That's because passmark is unreliable for that.

Use AIDA64 Cache & Memory benchmark.

Just use 3400 CL15 if you can't get 3600MHz CL16...

 

Also what kit do you use?

I managed to overclock (on a Zen+ CPU!) a Micron E die kit from 3200MHz 16-18-18-36 to 3733MHz 14-20-14-36:

1193097160_cachemem-3733CL14.thumb.png.c3872221fedec8c439955ef669837ffe.png

well.... trial version but daaaang your perf is better than mine where it shows

I reverted my sub timings to focus on speed for now.  I _really_ want to get this stuff running cl16 but every time I set 17 it just ignores that and if i set 16 it like literally boot loops.  can't even get into memtest
image.png.7025601d93220bb9293a2c9a77c83777.png

Link to post
Share on other sites
Posted
52 minutes ago, DaQuackers said:

well.... trial version but daaaang your perf is better than mine where it shows

It only shows memory latency - which is heavily affected by timings.

Read,write and copy are all heavily affected by clock speed - so i expect it to be slightly higher than mine (the difference between 3733MHz to 3800MHz is very small)

55 minutes ago, DaQuackers said:

I reverted my sub timings to focus on speed for now.  I _really_ want to get this stuff running cl16 but every time I set 17 it just ignores that and if i set 16 it like literally boot loops.  can't even get into memtest

I recommend you to go with 3600MHz not just because of the timings,but also because of the frequency of the Infinity Fabric.

The frequency of the Infinity Fabric is really low (950MHz is slow and affects memory latency as well).

 

And using Thaiphoon Burnerwould be very helpful to find out what RAM you have and it's potential.

 

Thaiphoon Burner: http://www.softnology.biz/files/thphn167.zip

 

Guide for using Thaiphoon Burner: 

1.Open Thaiphoon Burner

2.Click the "Read" button

3.Choose a RAM stick from the list,It will start with "Read SPD on SMBus"

4.Click the "Export" button

5.Choose "Copy plain text to Clipboard"

6.Paste in your reply

7.Repeat the guide for each RAM stick you have installed.

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE RTX 3080 GAMING OC | 4x 8GB Micron Rev.E (D9VPP) 3800MHz 16-19-14-21-58
Link to post
Share on other sites
Posted
  • Author
20 minutes ago, Vishera said:

 

6.Paste in your reply

 

module 1
-------------------------------------------------------------
                         MEMORY MODULE
-------------------------------------------------------------
Manufacturer             : Corsair
Series                   : Vengeance RGB PRO
Part Number              : CMW32GX4M2Z3600C18
Serial Number            : Undefined
JEDEC DIMM Label         : 16GB 1Rx8 PC4-2666U-UA0-10
Architecture             : DDR4 SDRAM UDIMM
Speed Grade              : DDR4-2666U
Capacity                 : 16 GB (8 components)
Organization             : 2048M x64 (1 rank)
Register Manufacturer    : N/A
Register Model           : N/A
Manufacturing Date       : Undefined
Manufacturing Location   : Taiwan
Revision / Raw Card      : 0000h / A0 (8 layers)
-------------------------------------------------------------
                        DRAM COMPONENTS
-------------------------------------------------------------
Manufacturer             : Micron Technology
Part Number              : D9XPF (MT40A2G8VA-062E:B)
Package                  : Standard Monolithic 78-ball FBGA
Die Density / Count      : 16 Gb B-die (Z22A / 17 nm) / 1 die
Composition              : 2048Mb x8 (128Mb x8 x 16 banks)
Input Clock Frequency    : 1333 MHz (0.750 ns)
Minimum Timing Delays    : 18-18-18-43-61
Read Latencies Supported : 24T, 23T, 22T, 21T, 20T, 19T, 18T...
Supply Voltage           : 1.20 V
XMP Certified            : 1799 MHz / 18-22-22-42-64 / 1.35 V
XMP Extreme              : Not programmed
SPD Revision             : 1.0 / January 2014
XMP Revision             : 2.0 / December 2013
-------------------------------------------------------------
                         THERMAL SENSOR
-------------------------------------------------------------
Manufacturer             : OnSemi
Model                    : N34TS04
Revision                 : 30h
Sensor Status            : Enabled
EVENT Output Control     : Disabled
Temperature Accuracy     : B-Grade
Temperature Resolution   : 0.0625 °C (12-bit ADC)
Current Temperature      : 31.938 °C / 89.488 °F
Negative Measurements    : Supported
Interrupt Capability     : Supported
-------------------------------------------------------------
                         SOURCE SPD DUMP
-------------------------------------------------------------
000  23 10 0C 02 86 29 00 08 00 00 00 03 01 03 80 00
010  00 00 06 0C FF FF 03 00 6C 6C 6C 11 01 6E 30 11
020  F0 0A 20 08 00 A8 1E 30 2B 00 00 00 00 00 00 00
030  00 00 00 00 00 00 00 00 00 00 00 00 16 36 16 36
040  16 36 16 36 00 00 2B 0C 2B 0C 2B 0C 2B 0C 00 00
050  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
060  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
070  00 00 00 00 00 ED 00 00 F6 F6 F6 F6 00 00 0E 82
080  11 11 00 00 00 00 00 00 00 00 00 00 00 00 00 00
090  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0A0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0B0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0C0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0D0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0E0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0F0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 B6 58
100  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
110  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
120  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
130  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
140  02 9E 00 00 00 00 00 00 00 43 4D 57 33 32 47 58
150  34 4D 32 5A 33 36 30 30 43 31 38 00 00 00 80 2C
160  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
170  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
180  0C 4A 01 20 00 00 00 00 00 A3 00 00 05 FF FF 03
190  00 50 62 62 10 BA 1D 30 11 F0 0A 20 08 00 B0 20
1A0  2D 00 00 00 00 00 00 00 00 BF 94 CD E4 E4 FE BB
1B0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1C0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1D0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1E0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1F0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00


module 2
-------------------------------------------------------------
                         MEMORY MODULE
-------------------------------------------------------------
Manufacturer             : Corsair
Series                   : Vengeance RGB PRO
Part Number              : CMW32GX4M2Z3600C18
Serial Number            : Undefined
JEDEC DIMM Label         : 16GB 1Rx8 PC4-2666U-UA0-10
Architecture             : DDR4 SDRAM UDIMM
Speed Grade              : DDR4-2666U
Capacity                 : 16 GB (8 components)
Organization             : 2048M x64 (1 rank)
Register Manufacturer    : N/A
Register Model           : N/A
Manufacturing Date       : Undefined
Manufacturing Location   : Taiwan
Revision / Raw Card      : 0000h / A0 (8 layers)
-------------------------------------------------------------
                        DRAM COMPONENTS
-------------------------------------------------------------
Manufacturer             : Micron Technology
Part Number              : D9XPF (MT40A2G8VA-062E:B)
Package                  : Standard Monolithic 78-ball FBGA
Die Density / Count      : 16 Gb B-die (Z22A / 17 nm) / 1 die
Composition              : 2048Mb x8 (128Mb x8 x 16 banks)
Input Clock Frequency    : 1333 MHz (0.750 ns)
Minimum Timing Delays    : 18-18-18-43-61
Read Latencies Supported : 24T, 23T, 22T, 21T, 20T, 19T, 18T...
Supply Voltage           : 1.20 V
XMP Certified            : 1799 MHz / 18-22-22-42-64 / 1.35 V
XMP Extreme              : Not programmed
SPD Revision             : 1.0 / January 2014
XMP Revision             : 2.0 / December 2013
-------------------------------------------------------------
                         THERMAL SENSOR
-------------------------------------------------------------
Manufacturer             : OnSemi
Model                    : N34TS04
Revision                 : 30h
Sensor Status            : Enabled
EVENT Output Control     : Disabled
Temperature Accuracy     : B-Grade
Temperature Resolution   : 0.0625 °C (12-bit ADC)
Current Temperature      : 34.563 °C / 94.213 °F
Negative Measurements    : Supported
Interrupt Capability     : Supported
-------------------------------------------------------------
                         SOURCE SPD DUMP
-------------------------------------------------------------
000  23 10 0C 02 86 29 00 08 00 00 00 03 01 03 80 00
010  00 00 06 0C FF FF 03 00 6C 6C 6C 11 01 6E 30 11
020  F0 0A 20 08 00 A8 1E 30 2B 00 00 00 00 00 00 00
030  00 00 00 00 00 00 00 00 00 00 00 00 16 36 16 36
040  16 36 16 36 00 00 2B 0C 2B 0C 2B 0C 2B 0C 00 00
050  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
060  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
070  00 00 00 00 00 ED 00 00 F6 F6 F6 F6 00 00 0E 82
080  11 11 00 00 00 00 00 00 00 00 00 00 00 00 00 00
090  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0A0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0B0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0C0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0D0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0E0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0F0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 B6 58
100  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
110  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
120  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
130  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
140  02 9E 00 00 00 00 00 00 00 43 4D 57 33 32 47 58
150  34 4D 32 5A 33 36 30 30 43 31 38 00 00 00 80 2C
160  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
170  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
180  0C 4A 01 20 00 00 00 00 00 A3 00 00 05 FF FF 03
190  00 50 62 62 10 BA 1D 30 11 F0 0A 20 08 00 B0 20
1A0  2D 00 00 00 00 00 00 00 00 BF 94 CD E4 E4 FE BB
1B0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1C0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1D0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1E0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
1F0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

Link to post
Share on other sites
Posted

So you have:

Micron B die 2GB x8

Single rank

 

I don't know much about Micron B die,

But maybe @ShrimpBrime or someone else knows and can help you with that.

 

23 minutes ago, DaQuackers said:

not really sure what a lot of that data dump means but I'm excited to learn

I think that the SOURCE SPD DUMP is just JEDEC specifications that those modules support,for example:

FREQUENCY CAS RCD RP RAS RC FAW RRDS RRDL WR WTRS
1200 MHz 20 16 16 39 55 26 4 6 18 3
1200 MHz 18 16 16 39 55 26 4 6 18 3
1200 MHz 16 16 16 39 55 26 4 6 18 3
1067 MHz 15 15 15 35 49 23 4 6 16 3
933 MHz 14 13 13 30 43 20 4 5 14 3
933 MHz 13 13 13 30 43 20 4 5 14 3
800 MHz 12 11 11 26 37 17 3 4 12 2
800 MHz 11 11 11 26 37 17 3 4 12 2
667 MHz 10 9 9 22 31 14 3 4 10 2
667 MHz 9 9 9 22 31 14 3 4 10 2

 

 

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE RTX 3080 GAMING OC | 4x 8GB Micron Rev.E (D9VPP) 3800MHz 16-19-14-21-58
Link to post
Share on other sites
Posted
  • Solution

I see what happened.

 

In the screen shot shows Infinity Fabric (North Bridge according to AIDA64 benchmark) running a 1:2 divider or "uncoupled" if you'd like to put it that way. Some boards will automatically do this. Thus the loss in latency.

 

However....

 

If you loosen the timings with a 1:2 IF to MEM divider, you might be able to run 4000mhz and this will increase bandwidth and lower latency slightly. Though the latency will stay inherently higher until you've reached a higher frequency, maybe 4400mhz.

 

Generally the Micron is ok with a little extra voltage. 1.4v is ok. Past this I recommend actively cooling them. Fan right on the sticks. Cooler brings better stability. It's just as important as the timings.

 

So force the 1:1 divider, or go on an OC climb. CL20-24-24 with XMP enabled, memory training should be enabled as well. Not sure if that Gigabyte board has training algorithms though. Something that's nice about Asus's top end boards. You can really OC the memory, so the board does matter too. 

 

GLHF!

Link to post
Share on other sites
Posted
  • Author
1 minute ago, ShrimpBrime said:

I see what happened.

 

In the screen shot shows Infinity Fabric (North Bridge according to AIDA64 benchmark) running a 1:2 divider or "uncoupled" if you'd like to put it that way. Some boards will automatically do this. Thus the loss in latency.

 

However....

 

If you loosen the timings with a 1:2 IF to MEM divider, you might be able to run 4000mhz and this will increase bandwidth and lower latency slightly. Though the latency will stay inherently higher until you've reached a higher frequency, maybe 4400mhz.

 

Generally the Micron is ok with a little extra voltage. 1.4v is ok. Past this I recommend actively cooling them. Fan right on the sticks. Cooler brings better stability. It's just as important as the timings.

 

So force the 1:1 divider, or go on an OC climb. CL20-24-24 with XMP enabled, memory training should be enabled as well. Not sure if that Gigabyte board has training algorithms though. Something that's nice about Asus's top end boards. You can really OC the memory, so the board does matter too. 

 

GLHF!

Thanks for the advice!

 

The board is actually an Asus TUF x570 pro wifi, not sure why it shows as a gigabyte board.  Not exactly their top of the line but not a bad board either

 

I didnt realized it had auto-decoupled that makes a ton of sense though!

Link to post
Share on other sites
Posted
1 minute ago, DaQuackers said:

I didnt realized it had auto-decoupled that makes a ton of sense though!

I actually pointed that out but worded it badly 😄

2 hours ago, Vishera said:

I recommend you to go with 3600MHz not just because of the timings,but also because of the frequency of the Infinity Fabric.

The frequency of the Infinity Fabric is really low (950MHz is slow and affects memory latency as well).

Boards that do auto-decouple tend to do that at frequencies higher than 3600MHz.

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE RTX 3080 GAMING OC | 4x 8GB Micron Rev.E (D9VPP) 3800MHz 16-19-14-21-58
Link to post
Share on other sites
Posted
  • Author
12 minutes ago, Vishera said:

I actually pointed that out but worded it badly 😄

Boards that do auto-decouple tend to do that at frequencies higher than 3600MHz.

Well interestingly it's not seeming to care when I try to manually set it.

 

The bios HAS a slot for fabric frequency, but no matter what I set it to Aida shows it as 950.1MHz

 

It also wouldn't post when I set it to 1900 (1:2)

Link to post
Share on other sites
Posted
  • Author

@Vishera && @ShrimpBrime I figured out my issue.  Turns out this motherboard has not 1 but 3 places you can enter the fclk frequency and which it respects is based on where and how you set some other parameters that can ALSO be set in multiple places.

 

Got it working now

Link to post
Share on other sites
Posted
1 minute ago, DaQuackers said:

@Vishera && @ShrimpBrime I figured out my issue.  Turns out this motherboard has not 1 but 3 places you can enter the fclk frequency and which it respects is based on where and how you set some other parameters that can ALSO be set in multiple places.

 

Got it working now

I wonder how it will fare in the AIDA64 test 😄

Glad to see that you got it working.

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE RTX 3080 GAMING OC | 4x 8GB Micron Rev.E (D9VPP) 3800MHz 16-19-14-21-58
Link to post
Share on other sites
Posted
  • Author
1 minute ago, Vishera said:

I wonder how it will fare in the AIDA64 test 😄

Glad to see that you got it working.

about 10ns faster on the latency just running the kit at its stock speed (3600) with slightly tighter secondary timings 🤣

cachemem_2_12_22_225.png.12a697d47e090a35352ebae3f77f42b5.png

Link to post
Share on other sites
Posted
  • Author

well to wrap up here's my findings.

My fabric WILL not run at 1900mhz or higher, 1866.7 is the last point at which it will post.
Also it will not run decoupled at anything other than a clean integer division.  So I can run 1:1 with mbus and fclk or I can run 1:2 mbus fclk.  nothing else will post.

Looks like this is where I cap out
 

cachemem_2_12_22_300.png.8d307ed50d29efac95ae3f26e08ceb3f.png

Link to post
Share on other sites
Posted
10 minutes ago, DaQuackers said:

well to wrap up here's my findings.

My fabric WILL not run at 1900mhz or higher, 1866.7 is the last point at which it will post.
Also it will not run decoupled at anything other than a clean integer division.  So I can run 1:1 with mbus and fclk or I can run 1:2 mbus fclk.  nothing else will post.

Your BIOS is weird 😄

Mine won't let me run at 1:2 so i am 1:1 all the time.

 

3733MHz is a decent speed,but i would try getting the timings lower.

 

A PC Enthusiast since 2011
AMD Ryzen 7 5700X@4.65GHz | GIGABYTE RTX 3080 GAMING OC | 4x 8GB Micron Rev.E (D9VPP) 3800MHz 16-19-14-21-58
Link to post
Share on other sites
Posted
  • Author
41 minutes ago, Vishera said:

3733MHz is a decent speed,but i would try getting the timings lower.

 

 

Yeah this is a WEIRD bios 🙃

I'd love to get the cas latency down but see again total lack of posting.  Cl17 it ignores and goes to 18 instead.  16 black screens and the dram light on the board turns red.  There might be some amount of voltage I could push to get it working, but after 2.5 days of mukin about I'll take today's gains for now and enjoy some actual time playing 🤣

 

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share
Go to topic listing CPUs, Motherboards, and Memory

×