F5F Stay Refreshed Hardware Desktop Unable to retrieve my memory at 3200MHz, please assist.

Unable to retrieve my memory at 3200MHz, please assist.

Unable to retrieve my memory at 3200MHz, please assist.

Pages (2): Previous 1 2
D
Deapple
Junior Member
6
11-30-2017, 11:24 AM
#11
It's unclear if they were set at 3200MHz or changed to 2133MHz automatically. You may need to adjust timing manually since the settings are locked in auto mode.
D
Deapple
11-30-2017, 11:24 AM #11

It's unclear if they were set at 3200MHz or changed to 2133MHz automatically. You may need to adjust timing manually since the settings are locked in auto mode.

S
Sztylet
Member
62
11-30-2017, 12:16 PM
#12
It’s best to watch a video and review timing and voltage details before attempting to overclock RAM. Manual adjustments can be tricky, so consider a helpful tutorial that matches your motherboard type. If you’re unsure or uncomfortable doing this yourself, avoid it.
S
Sztylet
11-30-2017, 12:16 PM #12

It’s best to watch a video and review timing and voltage details before attempting to overclock RAM. Manual adjustments can be tricky, so consider a helpful tutorial that matches your motherboard type. If you’re unsure or uncomfortable doing this yourself, avoid it.

M
Mountain_Man8
Member
182
12-21-2017, 12:39 AM
#13
Review the different versions on those units. I purchased two sets roughly a month apart, each with a Samsung B-die (v4.x) that performed much better than expected, while the other had significantly higher speeds. The second set was a completely different Corsair model—Corsair switched it to a Hynix part. From a technical standpoint, it’s not outright dishonest; they still meet the same rated rates when used together. However, mixing them with the original Samsung B-die didn’t yield close performance. I’ve been discussing this extensively, and here are the key points you should know:

- You’ll need to query your memory using specific details (see the linked comments).
- The main options were:
1) Swap in more of the older Samsung B-die modules.
2) Downgrade to 16GB instead of 32GB.
3) Adjust clock speeds slightly.
4) Loosen timings for a compromise.
5) Replace everything, especially focusing on the timing settings.

I tried #3 first and found it risky, so I went with #5. Now my CPU runs at 32GB (2x16GB) at 3200MHz with tight timings—far beyond what I was paying for. It’s been over a decade since I relied on these parts, but this marks the end of their use as my go-to supplier. Any similar changes can have serious consequences. The last time I switched vendors was for OCZ DDR1 in an AMD build!
M
Mountain_Man8
12-21-2017, 12:39 AM #13

Review the different versions on those units. I purchased two sets roughly a month apart, each with a Samsung B-die (v4.x) that performed much better than expected, while the other had significantly higher speeds. The second set was a completely different Corsair model—Corsair switched it to a Hynix part. From a technical standpoint, it’s not outright dishonest; they still meet the same rated rates when used together. However, mixing them with the original Samsung B-die didn’t yield close performance. I’ve been discussing this extensively, and here are the key points you should know:

- You’ll need to query your memory using specific details (see the linked comments).
- The main options were:
1) Swap in more of the older Samsung B-die modules.
2) Downgrade to 16GB instead of 32GB.
3) Adjust clock speeds slightly.
4) Loosen timings for a compromise.
5) Replace everything, especially focusing on the timing settings.

I tried #3 first and found it risky, so I went with #5. Now my CPU runs at 32GB (2x16GB) at 3200MHz with tight timings—far beyond what I was paying for. It’s been over a decade since I relied on these parts, but this marks the end of their use as my go-to supplier. Any similar changes can have serious consequences. The last time I switched vendors was for OCZ DDR1 in an AMD build!

K
Kraakka2005
Member
109
12-23-2017, 05:38 PM
#14
I noticed a decrease in MHz after installing more RAM on my system. The Ryzen 5 1600x with 2x8GB Gskill Red Ripjaws 16-16-16-36.1 (Samsung) Version 1.0 Asrock AB350 Fatality Gaming K4 Samsung 850 EVO M.2 500gb NVMe, Windows 10 Pro. I added another pair of 2x8GB Gskill Red Ripjaws 16-16-16-36.1 (Samsung). My speeds fell to 931MHz. I didn’t turn on XMP during the original build and didn’t realize my RAM was underperforming when additional sticks were installed. Both units looked identical in specs. I chose extra RAM for video editing, so more capacity seemed crucial rather than speed. I didn’t notice the actual throttling happening. I activated XMP in UEFI but the machine wouldn’t start. The XMP 2.0 loop caused the system to freeze. After a hard reboot, speeds returned to 931MHz. Each stick was tested with MemTest and passed without issues. I suspected another issue because I use two machines connected via HDMI KVM switch. Recently, a new 32-inch monitor showed flickering on either channel after trying XMP 2.0. I read about a tool called thaiphoon and decided to test it. After uploading snapshots of both RAM configurations, they differed significantly. Running the first set in A2 and B2 achieved 3200MHz easily, with no monitor issues. I’ve tried other RAM combinations without problems, so luck likely played a role. These two sets were purchased nearly a year apart. I’d like to mention that thaiphoon is a handy utility for verifying memory specifications. It reinforces the idea that buying matched RAM modules is better than separate ones if you can afford it. I’ll check if adjusting timings or swapping memory slots helps. We’ll find out.
K
Kraakka2005
12-23-2017, 05:38 PM #14

I noticed a decrease in MHz after installing more RAM on my system. The Ryzen 5 1600x with 2x8GB Gskill Red Ripjaws 16-16-16-36.1 (Samsung) Version 1.0 Asrock AB350 Fatality Gaming K4 Samsung 850 EVO M.2 500gb NVMe, Windows 10 Pro. I added another pair of 2x8GB Gskill Red Ripjaws 16-16-16-36.1 (Samsung). My speeds fell to 931MHz. I didn’t turn on XMP during the original build and didn’t realize my RAM was underperforming when additional sticks were installed. Both units looked identical in specs. I chose extra RAM for video editing, so more capacity seemed crucial rather than speed. I didn’t notice the actual throttling happening. I activated XMP in UEFI but the machine wouldn’t start. The XMP 2.0 loop caused the system to freeze. After a hard reboot, speeds returned to 931MHz. Each stick was tested with MemTest and passed without issues. I suspected another issue because I use two machines connected via HDMI KVM switch. Recently, a new 32-inch monitor showed flickering on either channel after trying XMP 2.0. I read about a tool called thaiphoon and decided to test it. After uploading snapshots of both RAM configurations, they differed significantly. Running the first set in A2 and B2 achieved 3200MHz easily, with no monitor issues. I’ve tried other RAM combinations without problems, so luck likely played a role. These two sets were purchased nearly a year apart. I’d like to mention that thaiphoon is a handy utility for verifying memory specifications. It reinforces the idea that buying matched RAM modules is better than separate ones if you can afford it. I’ll check if adjusting timings or swapping memory slots helps. We’ll find out.

Pages (2): Previous 1 2