F5F Stay Refreshed Power Users Overclocking What are overclocking factors.

What are overclocking factors.

What are overclocking factors.

Pages (2): 1 2 Next
G
ganelido
Member
64
08-24-2016, 12:05 AM
#1
I'm working on increasing the clock speed of my MSI GTX 1060 OC1 with a 6GB VRAM GPU. So far, I've reached around 2100 MHz core clock and about 4500 MHz memory clock. I've installed Hyix V-RAM, which is designed for overclocking. However, I'm hitting a barrier at 4550 MHz without any thermal throttling (temperature stays at 70°C), and I can't go higher beyond that.
G
ganelido
08-24-2016, 12:05 AM #1

I'm working on increasing the clock speed of my MSI GTX 1060 OC1 with a 6GB VRAM GPU. So far, I've reached around 2100 MHz core clock and about 4500 MHz memory clock. I've installed Hyix V-RAM, which is designed for overclocking. However, I'm hitting a barrier at 4550 MHz without any thermal throttling (temperature stays at 70°C), and I can't go higher beyond that.

S
StyleTrick
Senior Member
744
08-24-2016, 03:23 PM
#2
Every GPU has a maximum for overclocking, some programs can exceed this but the performance will likely become unstable afterward. Even if the overclock remains stable, it's important to verify that you have sufficient power to support it. Running in an unstable overclock can lead to game crashes during play and may cause permanent harm to the GPU or motherboard, so it's wise to adhere to the limits. If you aim to maximize your GPU's performance, using MSI AFTERBURNER is recommended.
S
StyleTrick
08-24-2016, 03:23 PM #2

Every GPU has a maximum for overclocking, some programs can exceed this but the performance will likely become unstable afterward. Even if the overclock remains stable, it's important to verify that you have sufficient power to support it. Running in an unstable overclock can lead to game crashes during play and may cause permanent harm to the GPU or motherboard, so it's wise to adhere to the limits. If you aim to maximize your GPU's performance, using MSI AFTERBURNER is recommended.

C
CaptianMClain
Member
61
08-24-2016, 05:55 PM
#3
Every GPU has a maximum overclocking boundary, some tools let you exceed it, but the results tend to become unstable once past that point. Even if the overclock remains stable, you must ensure your power supply can handle it. Running an unstable overclock can lead to game crashes and potentially permanent harm to the GPU or motherboard, so it's wise to stay within safe limits. If you aim for maximum performance, MSI AFTERBURNER is a good choice. Thank you for your advice.
C
CaptianMClain
08-24-2016, 05:55 PM #3

Every GPU has a maximum overclocking boundary, some tools let you exceed it, but the results tend to become unstable once past that point. Even if the overclock remains stable, you must ensure your power supply can handle it. Running an unstable overclock can lead to game crashes and potentially permanent harm to the GPU or motherboard, so it's wise to stay within safe limits. If you aim for maximum performance, MSI AFTERBURNER is a good choice. Thank you for your advice.

B
BrunoZed
Member
121
08-24-2016, 06:06 PM
#4
On the site where you purchased your GPU/PC, the specifications list core clock:xxxx and core clock(boost):xxxx. The boost value should reach its maximum. I won’t claim it’s the absolute limit since sometimes exceeding it slightly still works. Based on my experience, going 50MHz higher than the boost value caused my graphics card to crash within 5 to 20 minutes of gameplay (the more demanding the games were, the quicker the failure).

Hope this helps, if you’d like further details about GPU/CPU overclocking, check out JayzTwoCents on YouTube—there’s a lot to learn there.
B
BrunoZed
08-24-2016, 06:06 PM #4

On the site where you purchased your GPU/PC, the specifications list core clock:xxxx and core clock(boost):xxxx. The boost value should reach its maximum. I won’t claim it’s the absolute limit since sometimes exceeding it slightly still works. Based on my experience, going 50MHz higher than the boost value caused my graphics card to crash within 5 to 20 minutes of gameplay (the more demanding the games were, the quicker the failure).

Hope this helps, if you’d like further details about GPU/CPU overclocking, check out JayzTwoCents on YouTube—there’s a lot to learn there.

F
FuzzyMug
Senior Member
476
08-24-2016, 11:50 PM
#5
somesh101 :
i trying to overclock my msi gtx 1060 oc1 6gb GPU
up untill now i achieved 2100 Mhz core clock and about 4500 memory clock.
i got hynix vram which is meant to be best in overclocking.
but there seems to be a wall at 4550 Mhz without any thermals (temp 70C) i can't go beyond this.
Sounds like thats what you GPU is going to give you. I think this is good i cant get my 1070 even close to that
F
FuzzyMug
08-24-2016, 11:50 PM #5

somesh101 :
i trying to overclock my msi gtx 1060 oc1 6gb GPU
up untill now i achieved 2100 Mhz core clock and about 4500 memory clock.
i got hynix vram which is meant to be best in overclocking.
but there seems to be a wall at 4550 Mhz without any thermals (temp 70C) i can't go beyond this.
Sounds like thats what you GPU is going to give you. I think this is good i cant get my 1070 even close to that

I
imTri
Posting Freak
786
09-09-2016, 09:26 PM
#6
i'm trying to push overclock my msi gtx 1060 oc1 6gb graphics card. so far i've reached 2100 mhz core clock and around 4500 memory clock. i installed hyix vram which is designed for top overclocking performance. however, there appears to be a ceiling at 4550 mhz without any thermal throttling (temperature stays at 70c). it looks like that's what the card will ultimately deliver. i feel like this is impressive compared to my 1070, which barely reaches close to that level. sometimes it feels like i've won the silicon lottery, though not really. but you do have more cuda cores at the same power.
I
imTri
09-09-2016, 09:26 PM #6

i'm trying to push overclock my msi gtx 1060 oc1 6gb graphics card. so far i've reached 2100 mhz core clock and around 4500 memory clock. i installed hyix vram which is designed for top overclocking performance. however, there appears to be a ceiling at 4550 mhz without any thermal throttling (temperature stays at 70c). it looks like that's what the card will ultimately deliver. i feel like this is impressive compared to my 1070, which barely reaches close to that level. sometimes it feels like i've won the silicon lottery, though not really. but you do have more cuda cores at the same power.

S
Sheer_Lucke
Junior Member
48
09-10-2016, 05:14 AM
#7
Power constraints
Power output capabilities
Voltage supply options
Environmental conditions
Component temperatures
Overclocking expertise
Chance in the lottery
These factors all contribute greatly to successful overclocking. Including roughly a billion minor details would perfect the process.
Samsung memory performs best for this task. Hynix and Elpida are acceptable alternatives.
S
Sheer_Lucke
09-10-2016, 05:14 AM #7

Power constraints
Power output capabilities
Voltage supply options
Environmental conditions
Component temperatures
Overclocking expertise
Chance in the lottery
These factors all contribute greatly to successful overclocking. Including roughly a billion minor details would perfect the process.
Samsung memory performs best for this task. Hynix and Elpida are acceptable alternatives.

P
pipstick2004
Junior Member
17
09-30-2016, 05:08 PM
#8
Vellinious :
Power constraints
Power output
Voltage supply
Surrounding temperatures
Core temperature
Coolant temperature
Overclocking ability
Chance in the lottery
These factors all matter a lot when trying to overclock. Adding roughly a billion minor details gets you close.
Samsung memory performs best for this task. Hynix and Elpida work alright.
expanding your list ---
power limits -> can the GPU handle it?
power delivery -> what power can the PSU provide.
voltage - as mentioned before.
ambient - I understand.
coolant - if I could afford it, I'd buy a gaming X instead of this GPU's OC version.
😀
by the way, liquid cooling?
skil -> I'm just starting out
luck in lottery... still learning.
for around 4k less, I managed 1932 cores. 4492 memory instead of 2100 cores and 4800 memory.
I've seen somewhere that Hynix is top for OC, OCSAM are decent mem but they don't overclock well.
P
pipstick2004
09-30-2016, 05:08 PM #8

Vellinious :
Power constraints
Power output
Voltage supply
Surrounding temperatures
Core temperature
Coolant temperature
Overclocking ability
Chance in the lottery
These factors all matter a lot when trying to overclock. Adding roughly a billion minor details gets you close.
Samsung memory performs best for this task. Hynix and Elpida work alright.
expanding your list ---
power limits -> can the GPU handle it?
power delivery -> what power can the PSU provide.
voltage - as mentioned before.
ambient - I understand.
coolant - if I could afford it, I'd buy a gaming X instead of this GPU's OC version.
😀
by the way, liquid cooling?
skil -> I'm just starting out
luck in lottery... still learning.
for around 4k less, I managed 1932 cores. 4492 memory instead of 2100 cores and 4800 memory.
I've seen somewhere that Hynix is top for OC, OCSAM are decent mem but they don't overclock well.

S
Sly202001
Member
169
09-30-2016, 08:10 PM
#9
Interesting points here. These elements all contribute greatly to successful overclocking. Including roughly a billion minor details helps fine-tune the process. Samsung memory performs best, while Hynix and Elpida are decent alternatives.

Expanding on the terms:
power limits refer to whether the GPU can actually handle the power being applied.
Power delivery describes the PSU's capacity to supply consistent power.
Voltage availability depends on reference points mentioned earlier.
Ambient, coolant, and core temperatures all matter.
Core temps are crucial for stability.
Skill level affects how well you manage these variables.
Luck plays a role in finding the right balance.

For example, with around 4K less, I managed to hit 1932 cores. Using 4492 mem instead of 2100 and 4800 mem improved performance.
Some sources say Hynix is ideal for OC, though they’re not perfect.
Power limits are defined in the GPU BIOS; typical PSU outputs are about 200W, while custom boards can reach up to 350W. If the GPU hits its set power cap, it will throttle.
Power delivery quality matters too—stable voltage, minimal fluctuations, and a solid VRM are essential.
Coolant temperature is important; lower temps mean better performance.
With Pascal GPUs, cooler cores run faster and more efficiently.
I’ve read that Hynix offers good memory but struggles with overclocking.
Overall, it’s about balancing these factors carefully.
S
Sly202001
09-30-2016, 08:10 PM #9

Interesting points here. These elements all contribute greatly to successful overclocking. Including roughly a billion minor details helps fine-tune the process. Samsung memory performs best, while Hynix and Elpida are decent alternatives.

Expanding on the terms:
power limits refer to whether the GPU can actually handle the power being applied.
Power delivery describes the PSU's capacity to supply consistent power.
Voltage availability depends on reference points mentioned earlier.
Ambient, coolant, and core temperatures all matter.
Core temps are crucial for stability.
Skill level affects how well you manage these variables.
Luck plays a role in finding the right balance.

For example, with around 4K less, I managed to hit 1932 cores. Using 4492 mem instead of 2100 and 4800 mem improved performance.
Some sources say Hynix is ideal for OC, though they’re not perfect.
Power limits are defined in the GPU BIOS; typical PSU outputs are about 200W, while custom boards can reach up to 350W. If the GPU hits its set power cap, it will throttle.
Power delivery quality matters too—stable voltage, minimal fluctuations, and a solid VRM are essential.
Coolant temperature is important; lower temps mean better performance.
With Pascal GPUs, cooler cores run faster and more efficiently.
I’ve read that Hynix offers good memory but struggles with overclocking.
Overall, it’s about balancing these factors carefully.

X
220
10-01-2016, 02:34 AM
#10
Somewhat interesting:
Power constraints
Power output
Voltage supply
Environmental conditions
Core temperature
Coolant temperature
Overclocking expertise
Chance in the lottery
These aspects all contribute greatly to overclocking. Adding roughly a billion minor factors, you're really close.
Samsung memory performs best for this purpose. Hynix and Elpida are acceptable.
Expanding the vocabulary:
power limits → can the GPU power handle it?
power delivery → what power capacity does the PSU provide.
voltage – reference from above.
ambient temperature – I understand.
coolant – if I could afford it, I’d buy a gaming X instead of this GPU’s OC version.
😀
By the way, regarding liquid cooling?
skil is not experienced.
luck in lottery… still learning.
For around 4k less, I achieved 1932 cores. 4492 memory instead of 2100 cores and 4800 memory.
I’ve seen that Hynix is top for OC Samsung, while OC SAMUN are decent mem but poor overclocking.
power limits → does the GPU handle the power? → The limit is defined in the GPU BIOS. Boards typically sit around 200 watts. Custom builds can hit up to 350 watts. If your GPU hits the set power cap in the BIOS, it will throttle.
power delivery → how much power the PSU can supply. → This is part of the picture. Also, the quality of the power supply matters… voltage variations, waveform consistency, etc. Power delivery depends on the VRM on the GPU itself.
voltage – reference above. → Availability of voltage. Many elements here. VRM temps, core temps, overall temps affect VDROOP. Heat loss to the core reduces efficiency. Also, is the GPU voltage restricted? Most Pascal GPUs (except a few LN2 models) cap at 1.093v. This isn’t critical unless you dive into sub-ambient cooling.
ambient / coolant / core → with Pascal, the cooler the better. The lower the core temperature, the higher the boost and the smoother the performance when it reaches it.
I’ve read that Hynix is best for OC Samsung, while OC SAMUN are solid mem but not great for overclocking.
power limits → power GPU can manage? → Check the GPU BIOS settings. Standard boards usually max out at 200 watts. Custom builds can reach up to 350 watts. When your GPU hits the prescribed limit in the BIOS, it will automatically throttle.
There’s a lot of uncertainty about the GPU itself. There’s limited information online. Most discussions focus on gaming X or higher models.
I’m not sure if I should update the GPU BIOS yet.
But the NVIDIA-SMI on Linux indicates 140W as the maximum, which is lower than the normal 120W, and one more thing I forgot.
I’m running the card at 105 watts.
I’m considering this GPU for a second card, so it should be sufficient (GX550).
Here’s the Cooler Master GX 550 Strorm edition.
I’m not exploring sub-ambient cooling just yet; I’m only using the included heatsink and fans.
The only adjustment I can make is the core voltage.
I’m still figuring out the details.
This is the only link I found, and I don’t have anything comparable.
X
xXStrikeBackXx
10-01-2016, 02:34 AM #10

Somewhat interesting:
Power constraints
Power output
Voltage supply
Environmental conditions
Core temperature
Coolant temperature
Overclocking expertise
Chance in the lottery
These aspects all contribute greatly to overclocking. Adding roughly a billion minor factors, you're really close.
Samsung memory performs best for this purpose. Hynix and Elpida are acceptable.
Expanding the vocabulary:
power limits → can the GPU power handle it?
power delivery → what power capacity does the PSU provide.
voltage – reference from above.
ambient temperature – I understand.
coolant – if I could afford it, I’d buy a gaming X instead of this GPU’s OC version.
😀
By the way, regarding liquid cooling?
skil is not experienced.
luck in lottery… still learning.
For around 4k less, I achieved 1932 cores. 4492 memory instead of 2100 cores and 4800 memory.
I’ve seen that Hynix is top for OC Samsung, while OC SAMUN are decent mem but poor overclocking.
power limits → does the GPU handle the power? → The limit is defined in the GPU BIOS. Boards typically sit around 200 watts. Custom builds can hit up to 350 watts. If your GPU hits the set power cap in the BIOS, it will throttle.
power delivery → how much power the PSU can supply. → This is part of the picture. Also, the quality of the power supply matters… voltage variations, waveform consistency, etc. Power delivery depends on the VRM on the GPU itself.
voltage – reference above. → Availability of voltage. Many elements here. VRM temps, core temps, overall temps affect VDROOP. Heat loss to the core reduces efficiency. Also, is the GPU voltage restricted? Most Pascal GPUs (except a few LN2 models) cap at 1.093v. This isn’t critical unless you dive into sub-ambient cooling.
ambient / coolant / core → with Pascal, the cooler the better. The lower the core temperature, the higher the boost and the smoother the performance when it reaches it.
I’ve read that Hynix is best for OC Samsung, while OC SAMUN are solid mem but not great for overclocking.
power limits → power GPU can manage? → Check the GPU BIOS settings. Standard boards usually max out at 200 watts. Custom builds can reach up to 350 watts. When your GPU hits the prescribed limit in the BIOS, it will automatically throttle.
There’s a lot of uncertainty about the GPU itself. There’s limited information online. Most discussions focus on gaming X or higher models.
I’m not sure if I should update the GPU BIOS yet.
But the NVIDIA-SMI on Linux indicates 140W as the maximum, which is lower than the normal 120W, and one more thing I forgot.
I’m running the card at 105 watts.
I’m considering this GPU for a second card, so it should be sufficient (GX550).
Here’s the Cooler Master GX 550 Strorm edition.
I’m not exploring sub-ambient cooling just yet; I’m only using the included heatsink and fans.
The only adjustment I can make is the core voltage.
I’m still figuring out the details.
This is the only link I found, and I don’t have anything comparable.

Pages (2): 1 2 Next