Jump to content
Search In
  • More options...
Find results that contain...
Find results in...

What is the best way to test ram stability

MingLee420
 Share

what is the best way to test overclocked ram stability apart from playing my favourite triple a games

  • CPU
    10700k 5ghz all core
    1.275, 0 Avx 47 Ring Ratio 
  • Motherboard
    MSI Z590 A Pro
  • RAM
    Kingston Fury 2x16gb 3600mhz
  • GPU
    RTX 3080 Gaming Oc Revision 2
  • Case
    CoolerMaster H500
    2x 200mm front, 200mm top, a12x15 back
  • Storage
    970 Evo Plus 500gb, Sn550 1tb, 2tb barracuda, 860 evo 500gb, wd 5400rpm 1tb
  • PSU
    RM750x
Link to comment
Share on other sites

Link to post
Share on other sites

Memtest86

Prime95 memory test

Windows Memory Diagnostic

Is it plugged in? Is it turned on? Are you sure? No, really. 

List of my systems:

Main PC:

Spoiler

HP z620 workstation, 2x Xeon e5-2690 v0, EVGA GTX 980 SC, 8x4gb DDR3-1333, 2x 2tb HDD, 1tb boot SATA SSD, Windows 10 Pro

Linux system:

Spoiler

ProDesk 600 g1 SFF, i5-4590, Quadro k620, 2x4gb DDR3-1333, 64gb boot SSD, 250gb storage HDD from the ewaste pile with many bad sectors, Ubuntu 20.04 LTS

XP system:

Spoiler

Compaq DC7600 SFF, Pentium 4 HT 631, MSI GT 210 passive, 2x512m DDR2-667, 80gb Spinmaster HDD

Testbench:

Spoiler

Optiplex 760 board, GameMax VP-600 PSU, 4x 64gb AliExpress SSD (Windows XP, Vista, 7, 10), 1tb HDD (storage)

(currently has a Core 2 Quad Q9550 and 4x2gb DDR2-800 but I change its config frequently)

 

Link to comment
Share on other sites

Link to post
Share on other sites

Mel0nMan pretty much provided the best answer. I just wanna recommend MemTest86 aswell 🙂

MOBO: ASUS ROG STRIX Z370-F Gaming   CPU: Intel Core i7-8700K@4.7GHz, -0.095mV (Custom IHS, LM)   GPU: ASUS ROG STRIX RTX 3080 Ti OC LC@2130MHz + 20Gbps MEM OC 

RAM: G.SKILL TridentZ RGB 4x8GB@3700MHz, 16-16-16-36, 1.35V   NVMe SSD: Samsung 970 EVO Plus 1TB   #1 SATA SSD: Crucial MX500 2TB   #2 SATA SSD: Crucial MX300 1.05TB

PSU: Corsair HXi 1200W 80-PLUS Platinum   CASE: Corsair iCUE 7000X RGB   MONITOR: Samsung 49" Odyssey CRG9 QLED   SPEAKERS: Corsair SP2500 + FX Audio DAC-X6

 

Full specs and setup on my profile-page

Link to comment
Share on other sites

Link to post
Share on other sites

These days I have been using TM5, the Anta Extreme profile. I also like Linpack Xtreme, I use v1.15.

AMD R9 5900X | Thermalright Frost Commander 140/ TY-143
Asus Strix B550-F Gaming | 4x8GB G.Skill Trident Z 3200C14
EVGA RTX 3070Ti FTW3 Ultra | WD SN750 500GB, 1TB

Seasonic Prime GX-750 | Fractal Design Meshify C | iPPC 3K

Link to comment
Share on other sites

Link to post
Share on other sites

15 hours ago, Mel0nMan said:

Memtest86

Prime95 memory test

Windows Memory Diagnostic

 

15 hours ago, BetteBalterZen said:

Mel0nMan pretty much provided the best answer. I just wanna recommend MemTest86 aswell 🙂

 

14 hours ago, freeagent said:

These days I have been using TM5, the Anta Extreme profile. I also like Linpack Xtreme, I use v1.15.

i did memtest86 over night and passed, 0 errors 

i got 3600mhz c16-21-21-38 but i might tighten it more but if i try 3733mhz it just dosen't boot stock was c18-22-22-39

  • CPU
    10700k 5ghz all core
    1.275, 0 Avx 47 Ring Ratio 
  • Motherboard
    MSI Z590 A Pro
  • RAM
    Kingston Fury 2x16gb 3600mhz
  • GPU
    RTX 3080 Gaming Oc Revision 2
  • Case
    CoolerMaster H500
    2x 200mm front, 200mm top, a12x15 back
  • Storage
    970 Evo Plus 500gb, Sn550 1tb, 2tb barracuda, 860 evo 500gb, wd 5400rpm 1tb
  • PSU
    RM750x
Link to comment
Share on other sites

Link to post
Share on other sites

memtest86 is not to test stability, its to test defects.

CPU: i7-8700k 5GHz@1.25v | MoBo: Asus ROG Maximus X Formula | RAM: G.Skill F4-3600C15D-16GTZ @4000CL17GPU: RTX 2080Ti | PSU: Evga Supernova G2 750w | 

Case: Lian Li 011D XL | Storage: Samsung 970 EVO M.2 NVMe 500GB, Crucial MX500 500GB | Soundcard: Soundblaster ZXR | Mouse: Razer Viper Mini | Keyboard: Razer Huntsman TE Monitor: DELL AW2521H @360Hz |

 

Link to comment
Share on other sites

Link to post
Share on other sites

13 minutes ago, DarkSmith2 said:

memtest86 is not to test stability, its to test defects.

should i do prime95 and or hci memtest?

  • CPU
    10700k 5ghz all core
    1.275, 0 Avx 47 Ring Ratio 
  • Motherboard
    MSI Z590 A Pro
  • RAM
    Kingston Fury 2x16gb 3600mhz
  • GPU
    RTX 3080 Gaming Oc Revision 2
  • Case
    CoolerMaster H500
    2x 200mm front, 200mm top, a12x15 back
  • Storage
    970 Evo Plus 500gb, Sn550 1tb, 2tb barracuda, 860 evo 500gb, wd 5400rpm 1tb
  • PSU
    RM750x
Link to comment
Share on other sites

Link to post
Share on other sites

nvm 

Mod please delete

MOBO: ASUS ROG STRIX Z370-F Gaming   CPU: Intel Core i7-8700K@4.7GHz, -0.095mV (Custom IHS, LM)   GPU: ASUS ROG STRIX RTX 3080 Ti OC LC@2130MHz + 20Gbps MEM OC 

RAM: G.SKILL TridentZ RGB 4x8GB@3700MHz, 16-16-16-36, 1.35V   NVMe SSD: Samsung 970 EVO Plus 1TB   #1 SATA SSD: Crucial MX500 2TB   #2 SATA SSD: Crucial MX300 1.05TB

PSU: Corsair HXi 1200W 80-PLUS Platinum   CASE: Corsair iCUE 7000X RGB   MONITOR: Samsung 49" Odyssey CRG9 QLED   SPEAKERS: Corsair SP2500 + FX Audio DAC-X6

 

Full specs and setup on my profile-page

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, DarkSmith2 said:

memtest86 is not to test stability, its to test defects.

How is it different? 

RAM OC'ed too hard causing errors would be the same as defect RAM, wouldn't it? In terms of errors produced

MOBO: ASUS ROG STRIX Z370-F Gaming   CPU: Intel Core i7-8700K@4.7GHz, -0.095mV (Custom IHS, LM)   GPU: ASUS ROG STRIX RTX 3080 Ti OC LC@2130MHz + 20Gbps MEM OC 

RAM: G.SKILL TridentZ RGB 4x8GB@3700MHz, 16-16-16-36, 1.35V   NVMe SSD: Samsung 970 EVO Plus 1TB   #1 SATA SSD: Crucial MX500 2TB   #2 SATA SSD: Crucial MX300 1.05TB

PSU: Corsair HXi 1200W 80-PLUS Platinum   CASE: Corsair iCUE 7000X RGB   MONITOR: Samsung 49" Odyssey CRG9 QLED   SPEAKERS: Corsair SP2500 + FX Audio DAC-X6

 

Full specs and setup on my profile-page

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, BetteBalterZen said:

How is it different? 

RAM OC'ed too hard causing errors would be the same as defect RAM, wouldn't it? In terms of errors produced

I believe that's correct. Programs like memtest86 are just checking to make sure that the RAM can accurately store and retrieve data. I don't believe there is any meaningful difference between a mistake due to a hardware failure or RAM being pushed too far.

 

I'm interested to see if I'm wrong, though.

Link to comment
Share on other sites

Link to post
Share on other sites

5 hours ago, BetteBalterZen said:

RAM OC'ed too hard causing errors would be the same as defect RAM, wouldn't it? In terms of errors produced

Memtest86 is just unreliable to find errors/instability caused by overclocking. But its very good and fast to find any broken cells at stock settings.

My guess is that the Algorithms being used are just ineffective due to its age and relatively low load for todays Hard- and Software.

 

Or in other words, you can pass memtest86+ while HCI Memtest, Karhu, Testmem5 or Google Stress App will find errors.

So my answer is no, you wont necessarily find the same amount of errors unless a stick is broken.

CPU: i7-8700k 5GHz@1.25v | MoBo: Asus ROG Maximus X Formula | RAM: G.Skill F4-3600C15D-16GTZ @4000CL17GPU: RTX 2080Ti | PSU: Evga Supernova G2 750w | 

Case: Lian Li 011D XL | Storage: Samsung 970 EVO M.2 NVMe 500GB, Crucial MX500 500GB | Soundcard: Soundblaster ZXR | Mouse: Razer Viper Mini | Keyboard: Razer Huntsman TE Monitor: DELL AW2521H @360Hz |

 

Link to comment
Share on other sites

Link to post
Share on other sites

2 minutes ago, DarkSmith2 said:

Memtest86 is just unreliable to find errors/instability caused by overclocking. But its very good and fast to find any broken cells at stock settings.

My guess is that the Algorithms being used are just ineffective due to its age and relatively low load for todays Hard- and Software.

 

Or in other words, you can pass memtest86+ while HCI Memtest, Karhu, Testmem5 or Google Stress App will find errors.

So my answer is no, you wont necessarily find the same amount of errors unless a stick is broken.

Oh alright, gotcha. 

What RAM test would you use out of all the ones you mentioned? 

MOBO: ASUS ROG STRIX Z370-F Gaming   CPU: Intel Core i7-8700K@4.7GHz, -0.095mV (Custom IHS, LM)   GPU: ASUS ROG STRIX RTX 3080 Ti OC LC@2130MHz + 20Gbps MEM OC 

RAM: G.SKILL TridentZ RGB 4x8GB@3700MHz, 16-16-16-36, 1.35V   NVMe SSD: Samsung 970 EVO Plus 1TB   #1 SATA SSD: Crucial MX500 2TB   #2 SATA SSD: Crucial MX300 1.05TB

PSU: Corsair HXi 1200W 80-PLUS Platinum   CASE: Corsair iCUE 7000X RGB   MONITOR: Samsung 49" Odyssey CRG9 QLED   SPEAKERS: Corsair SP2500 + FX Audio DAC-X6

 

Full specs and setup on my profile-page

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, BetteBalterZen said:

Oh alright, gotcha. 

What RAM test would you use out of all the ones you mentioned? 

im using Karhu because its easy to use and "fast" only 2.5hours for 16GB.. but its not free. Some people would advice to use multiple stresstests, some even use Linpack in addition to any other stresstests to look after variations in performance for a glimpse of instability.

But tbh use whichever you like of those options. Just not Memtest86
 

CPU: i7-8700k 5GHz@1.25v | MoBo: Asus ROG Maximus X Formula | RAM: G.Skill F4-3600C15D-16GTZ @4000CL17GPU: RTX 2080Ti | PSU: Evga Supernova G2 750w | 

Case: Lian Li 011D XL | Storage: Samsung 970 EVO M.2 NVMe 500GB, Crucial MX500 500GB | Soundcard: Soundblaster ZXR | Mouse: Razer Viper Mini | Keyboard: Razer Huntsman TE Monitor: DELL AW2521H @360Hz |

 

Link to comment
Share on other sites

Link to post
Share on other sites

7 minutes ago, DarkSmith2 said:

im using Karhu because its easy to use and "fast" only 2.5hours for 16GB.. but its not free. Some people would advice to use multiple stresstests, some even use Linpack in addition to any other stresstests to look after variations in performance for a glimpse of instability.

But tbh use whichever you like of those options. Just not Memtest86
 

Cool, thank you 🙂

MOBO: ASUS ROG STRIX Z370-F Gaming   CPU: Intel Core i7-8700K@4.7GHz, -0.095mV (Custom IHS, LM)   GPU: ASUS ROG STRIX RTX 3080 Ti OC LC@2130MHz + 20Gbps MEM OC 

RAM: G.SKILL TridentZ RGB 4x8GB@3700MHz, 16-16-16-36, 1.35V   NVMe SSD: Samsung 970 EVO Plus 1TB   #1 SATA SSD: Crucial MX500 2TB   #2 SATA SSD: Crucial MX300 1.05TB

PSU: Corsair HXi 1200W 80-PLUS Platinum   CASE: Corsair iCUE 7000X RGB   MONITOR: Samsung 49" Odyssey CRG9 QLED   SPEAKERS: Corsair SP2500 + FX Audio DAC-X6

 

Full specs and setup on my profile-page

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, DarkSmith2 said:

Memtest86 is just unreliable to find errors/instability caused by overclocking. But its very good and fast to find any broken cells at stock settings.

My guess is that the Algorithms being used are just ineffective due to its age and relatively low load for todays Hard- and Software.

 

Or in other words, you can pass memtest86+ while HCI Memtest, Karhu, Testmem5 or Google Stress App will find errors.

So my answer is no, you wont necessarily find the same amount of errors unless a stick is broken.

ok ill try hci memtest

  • CPU
    10700k 5ghz all core
    1.275, 0 Avx 47 Ring Ratio 
  • Motherboard
    MSI Z590 A Pro
  • RAM
    Kingston Fury 2x16gb 3600mhz
  • GPU
    RTX 3080 Gaming Oc Revision 2
  • Case
    CoolerMaster H500
    2x 200mm front, 200mm top, a12x15 back
  • Storage
    970 Evo Plus 500gb, Sn550 1tb, 2tb barracuda, 860 evo 500gb, wd 5400rpm 1tb
  • PSU
    RM750x
Link to comment
Share on other sites

Link to post
Share on other sites

5 hours ago, YoungBlade said:

I believe that's correct. Programs like memtest86 are just checking to make sure that the RAM can accurately store and retrieve data. I don't believe there is any meaningful difference between a mistake due to a hardware failure or RAM being pushed too far.

 

I'm interested to see if I'm wrong, though.

You are correct. A bit-flip can be caused by both faulty silicon or improper timings and voltages.

 

3 hours ago, DarkSmith2 said:

Memtest86 is just unreliable to find errors/instability caused by overclocking. But its very good and fast to find any broken cells at stock settings.

My guess is that the Algorithms being used are just ineffective due to its age and relatively low load for todays Hard- and Software.

 

Or in other words, you can pass memtest86+ while HCI Memtest, Karhu, Testmem5 or Google Stress App will find errors.

So my answer is no, you wont necessarily find the same amount of errors unless a stick is broken.

Memtest86, and others like it, are better suited for finding specific errors in RAM where there's a specific hardware region at fault. This will manifest itself with a specific address showing up in error over and over.

 

However incorrect timings or voltages will cause errors to occur in random address space (more often than not). The problem here is that for every random error caught in this way, there's a whole lot more errors occurring that can't be identified in real-time as the test isn't checking that specific address at that exact specific time. The algorithm here isn't at fault because you can't test the entire stick of RAM at the exact same time that each and every error occurs from improper timings and voltages. Or to put it another way, testing for these types of errors and finding them is like showing the proverbial 'tip of the iceberg'; for every error found, there's a whole lot more occurring that's not being reported.

Link to comment
Share on other sites

Link to post
Share on other sites

15 hours ago, DarkSmith2 said:

im using Karhu because its easy to use and "fast" only 2.5hours for 16GB.. but its not free. Some people would advice to use multiple stresstests, some even use Linpack in addition to any other stresstests to look after variations in performance for a glimpse of instability.

But tbh use whichever you like of those options. Just not Memtest86
 

 I have 16 processes at 2048mb overnight some of them are like above 500 percent and some are under 400 and 200 with a few at a 1000 percent do you think tjata good enough

16379874804972510442074771741447.jpg

  • CPU
    10700k 5ghz all core
    1.275, 0 Avx 47 Ring Ratio 
  • Motherboard
    MSI Z590 A Pro
  • RAM
    Kingston Fury 2x16gb 3600mhz
  • GPU
    RTX 3080 Gaming Oc Revision 2
  • Case
    CoolerMaster H500
    2x 200mm front, 200mm top, a12x15 back
  • Storage
    970 Evo Plus 500gb, Sn550 1tb, 2tb barracuda, 860 evo 500gb, wd 5400rpm 1tb
  • PSU
    RM750x
Link to comment
Share on other sites

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


×