F5F Stay Refreshed Power Users Overclocking Adjusting the speed of TridentZ RGB to 3600Mhz with settings 17-18-18-38

Adjusting the speed of TridentZ RGB to 3600Mhz with settings 17-18-18-38

Adjusting the speed of TridentZ RGB to 3600Mhz with settings 17-18-18-38

Pages (2): 1 2 Next
D
DarkKnight_BR
Junior Member
4
03-02-2016, 03:46 PM
#1
Hi all, I recently tried overclocking my RAM and got two solid results without changing other settings. The speeds I achieved were 16-16-16-36 3600Mhz and 17-17-17-37 3866Mhz. I have two questions: 1) is it safe to run RAMs at these speeds? and 2) do you know any better settings I can use?
D
DarkKnight_BR
03-02-2016, 03:46 PM #1

Hi all, I recently tried overclocking my RAM and got two solid results without changing other settings. The speeds I achieved were 16-16-16-36 3600Mhz and 17-17-17-37 3866Mhz. I have two questions: 1) is it safe to run RAMs at these speeds? and 2) do you know any better settings I can use?

F
Firamir
Junior Member
26
03-15-2016, 04:03 PM
#2
The goal is to minimize the actual latency as much as possible. Raising the voltage can help avoid boot issues, but there are two issues to consider:
(1) Risk of harming memory chips.
(2) Potential damage to the CPU's memory controller.
1.4 volts represents a slight increase of only 3.7% compared to 1.35. I've tested these modules at 1.42, which is roughly a 5.2% overvoltage. If you aim to boot and run 3733 at 16 to achieve a lower true latency of 8.57nS, then consider using 1.4 volts, which remains within the permitted +5% deviation.
Keep in mind the impact of memory overclocking; it doesn't significantly improve overall performance. Unless you're using a specialized application or...
F
Firamir
03-15-2016, 04:03 PM #2

The goal is to minimize the actual latency as much as possible. Raising the voltage can help avoid boot issues, but there are two issues to consider:
(1) Risk of harming memory chips.
(2) Potential damage to the CPU's memory controller.
1.4 volts represents a slight increase of only 3.7% compared to 1.35. I've tested these modules at 1.42, which is roughly a 5.2% overvoltage. If you aim to boot and run 3733 at 16 to achieve a lower true latency of 8.57nS, then consider using 1.4 volts, which remains within the permitted +5% deviation.
Keep in mind the impact of memory overclocking; it doesn't significantly improve overall performance. Unless you're using a specialized application or...

R
ReborntoKill
Posting Freak
821
03-30-2016, 02:46 AM
#3
just keep an eye on the temperatures. like did you increase the voltage (if you did, that might be concerning) .. and just a note, depending on your other configuration, it could run slower than at the optimal settings at 3600Mhz. you should perform some benchmarks to see which is actually more efficient.
R
ReborntoKill
03-30-2016, 02:46 AM #3

just keep an eye on the temperatures. like did you increase the voltage (if you did, that might be concerning) .. and just a note, depending on your other configuration, it could run slower than at the optimal settings at 3600Mhz. you should perform some benchmarks to see which is actually more efficient.

A
alan0621
Member
212
04-04-2016, 01:02 AM
#4
Doctor Rob is suggesting you monitor the temperatures closely. He mentions that increasing the voltage could be concerning and advises running benchmarks to compare performance at different settings. He also notes he's a beginner in overclocking and is looking for guidance on adjusting timings and frequency. It seems he wants to explore whether lowering timings with higher frequencies might improve results.
A
alan0621
04-04-2016, 01:02 AM #4

Doctor Rob is suggesting you monitor the temperatures closely. He mentions that increasing the voltage could be concerning and advises running benchmarks to compare performance at different settings. He also notes he's a beginner in overclocking and is looking for guidance on adjusting timings and frequency. It seems he wants to explore whether lowering timings with higher frequencies might improve results.

O
okA_
Member
141
04-09-2016, 09:33 AM
#5
I've adjusted several kits, such as G.Skill Trident Z RGB 3200 @ 14-14-14-34 and Trident Z RGB 3600 @ 16-16-16-36. Both remained steady at 3733 @ 16-16-16-36. Have you experimented with 3733 @ 16-16-16-36?

To determine true latency in nanoseconds, the formula is:
3.6GHz (frequency) divided by 2 (DDR or Double Data Rate) equals 1.8GHz.
Taking the time constant of 1 and dividing by 1.8 gives approximately 0.5556.
Multiplying by 16 yields about 8.89 nanoseconds (true latency).

Values like 3.600 @ 16 equal 8.89, 3.866 @ 17 is 8.79, and so on. Only the top-tier and pricier kits can reach around 8 nanoseconds. As you approach that level, keeping stability becomes harder, often needing a voltage boost.

If your kit needs stable performance at 3733 with 16-16-16-36 timings, aiming for 8.57 nanoseconds would be ideal. My 3600 model could hit 4.0GHz @ 17-17-17-37 but needed a slight overvoltage of 1.38 to stay stable.

For more details, check Crucial's paper on speed versus latency: http://www.crucial.com/usa/en/memory-per...ed-latency
And use this simple latency calculator: https://notkyon.moe/ram-latency.htm
O
okA_
04-09-2016, 09:33 AM #5

I've adjusted several kits, such as G.Skill Trident Z RGB 3200 @ 14-14-14-34 and Trident Z RGB 3600 @ 16-16-16-36. Both remained steady at 3733 @ 16-16-16-36. Have you experimented with 3733 @ 16-16-16-36?

To determine true latency in nanoseconds, the formula is:
3.6GHz (frequency) divided by 2 (DDR or Double Data Rate) equals 1.8GHz.
Taking the time constant of 1 and dividing by 1.8 gives approximately 0.5556.
Multiplying by 16 yields about 8.89 nanoseconds (true latency).

Values like 3.600 @ 16 equal 8.89, 3.866 @ 17 is 8.79, and so on. Only the top-tier and pricier kits can reach around 8 nanoseconds. As you approach that level, keeping stability becomes harder, often needing a voltage boost.

If your kit needs stable performance at 3733 with 16-16-16-36 timings, aiming for 8.57 nanoseconds would be ideal. My 3600 model could hit 4.0GHz @ 17-17-17-37 but needed a slight overvoltage of 1.38 to stay stable.

For more details, check Crucial's paper on speed versus latency: http://www.crucial.com/usa/en/memory-per...ed-latency
And use this simple latency calculator: https://notkyon.moe/ram-latency.htm

P
prettykitty75
Junior Member
3
04-09-2016, 11:08 AM
#6
CompuTronix :
NewsFromHell
I've adjusted several kits, such as the G.Skill Trident Z RGB 3200 @ 14-14-14-34 and Trident Z RGB 3600 @ 16-16-16-36. Both remained stable at 3733 @ 16-16-16-36.
Have you experimented with 3733 @ 16-16-16-36?

To determine true latency in nanoseconds:
3.6GHz divided by 2 (DDR or Double Data Rate) equals 1.8GHz.
Taking the time constant (1) divided by 1.8 gives approximately 0.5556.
Multiplying by 16 yields about 8.89 nanoseconds (true latency).

Values like 3.600 @ 16 = 8.89, 3.866 @ 17 = 8.79, 3.200 @ 14 = 8.75, and 3.733 @ 16 = 8.57.
Only the top-tier and pricier kits can reach around 8 nanoseconds (nS). The nearer to 8ns, the harder it is to keep stability, eventually requiring a voltage boost.

If your kit can maintain stability at 3733 using 16-16-16-36 timings, then 8.57 nanoseconds should be preferable.
My 3600 model could hit 4.0GHz @ 17-17-17-37, but it needed a slight overvoltage of 1.38 to stay stable.

For more details, see Crucial's paper on speed versus latency:
http://www.crucial.com/usa/en/memory-per...ed-latency
And a straightforward latency calculator is available here:
https://notkyon.moe/ram-latency.htm
P
prettykitty75
04-09-2016, 11:08 AM #6

CompuTronix :
NewsFromHell
I've adjusted several kits, such as the G.Skill Trident Z RGB 3200 @ 14-14-14-34 and Trident Z RGB 3600 @ 16-16-16-36. Both remained stable at 3733 @ 16-16-16-36.
Have you experimented with 3733 @ 16-16-16-36?

To determine true latency in nanoseconds:
3.6GHz divided by 2 (DDR or Double Data Rate) equals 1.8GHz.
Taking the time constant (1) divided by 1.8 gives approximately 0.5556.
Multiplying by 16 yields about 8.89 nanoseconds (true latency).

Values like 3.600 @ 16 = 8.89, 3.866 @ 17 = 8.79, 3.200 @ 14 = 8.75, and 3.733 @ 16 = 8.57.
Only the top-tier and pricier kits can reach around 8 nanoseconds (nS). The nearer to 8ns, the harder it is to keep stability, eventually requiring a voltage boost.

If your kit can maintain stability at 3733 using 16-16-16-36 timings, then 8.57 nanoseconds should be preferable.
My 3600 model could hit 4.0GHz @ 17-17-17-37, but it needed a slight overvoltage of 1.38 to stay stable.

For more details, see Crucial's paper on speed versus latency:
http://www.crucial.com/usa/en/memory-per...ed-latency
And a straightforward latency calculator is available here:
https://notkyon.moe/ram-latency.htm

N
N3onkirby
Junior Member
14
04-09-2016, 02:44 PM
#7
Sorry, I was editing my answer when you posted. You might want to read it again. Basically, your only option is to increase memory voltage.
Which CPU and motherboard do you have?
N
N3onkirby
04-09-2016, 02:44 PM #7

Sorry, I was editing my answer when you posted. You might want to read it again. Basically, your only option is to increase memory voltage.
Which CPU and motherboard do you have?

R
Raranono
Junior Member
7
04-09-2016, 06:10 PM
#8
CompuTronix:
Apologies, I was adjusting my response when you shared it. You may wish to review it once more. In short, your main choice is to boost the memory voltage.
Which CPU and motherboard are you using?
8700k stock
Asus Z370-A Prime
P.S. The best result I've achieved is 3866Mhz 17-17-17-37
R
Raranono
04-09-2016, 06:10 PM #8

CompuTronix:
Apologies, I was adjusting my response when you shared it. You may wish to review it once more. In short, your main choice is to boost the memory voltage.
Which CPU and motherboard are you using?
8700k stock
Asus Z370-A Prime
P.S. The best result I've achieved is 3866Mhz 17-17-17-37

D
DBAlucard
Member
211
04-28-2016, 10:46 AM
#9
Update: memtest64 detected an issue on 3866Mhz 17-17-17-37. The current best result is 16-16-16-36. Any suggestions for improvement?
D
DBAlucard
04-28-2016, 10:46 AM #9

Update: memtest64 detected an issue on 3866Mhz 17-17-17-37. The current best result is 16-16-16-36. Any suggestions for improvement?

K
ksamp3
Member
60
04-30-2016, 07:43 AM
#10
The goal is to minimize the actual latency as much as possible. Raising the memory voltage helps resolve boot issues, but there are two issues to consider:
(1) Risk of harming memory chips.
(2) Potential damage to the CPU's memory controller.
1.4 volts represents a slight increase of around 3.7% compared to 1.35. I've tested these modules at 1.42, which is approximately a 5.2% overvoltage. If you aim to boot and run 3733 at 16 to achieve a lower true latency of 8.57nS, then consider using 1.4 volts, which remains within the permitted +5% deviation.
Keep in mind that any memory overclocking should be viewed carefully. It offers limited impact on overall system performance. Unless you're using a particular application or game where memory speed is crucial, focus your efforts on CPU and GPU overclocking instead.
Also, ensure you thoroughly test and confirm memory stability; neglecting this can lead to serious data loss or corruption. Each BSOD event poses a significant risk, so it's wise to back up your system completely before diving into overclocking.
K
ksamp3
04-30-2016, 07:43 AM #10

The goal is to minimize the actual latency as much as possible. Raising the memory voltage helps resolve boot issues, but there are two issues to consider:
(1) Risk of harming memory chips.
(2) Potential damage to the CPU's memory controller.
1.4 volts represents a slight increase of around 3.7% compared to 1.35. I've tested these modules at 1.42, which is approximately a 5.2% overvoltage. If you aim to boot and run 3733 at 16 to achieve a lower true latency of 8.57nS, then consider using 1.4 volts, which remains within the permitted +5% deviation.
Keep in mind that any memory overclocking should be viewed carefully. It offers limited impact on overall system performance. Unless you're using a particular application or game where memory speed is crucial, focus your efforts on CPU and GPU overclocking instead.
Also, ensure you thoroughly test and confirm memory stability; neglecting this can lead to serious data loss or corruption. Each BSOD event poses a significant risk, so it's wise to back up your system completely before diving into overclocking.

Pages (2): 1 2 Next