F5F Stay Refreshed Power Users Overclocking 8700k OC without deliding?

8700k OC without deliding?

8700k OC without deliding?

Pages (3): Previous 1 2 3 Next
M
MX_Flame
Member
171
12-28-2017, 04:48 AM
#11
I believe a tailored fan profile in Corsair Link is possible, though it's been quite some time since I last used it—it was quite unstable for me, but it's been over a year. The stock fans were removed and replaced with quieter Cougar models, which helped boost the RPM.

I can push the RPM significantly. My ML140 Pro fans run smoothly up to 56% speed, while the rest of my setup remains nearly silent. The only noticeable sound now is air movement. The CD drive, thanks to the ROG software I installed, has been causing issues all day—about four to five times louder than the rest of the rig.

I keep them mainly because Corsair makes excellent fans and I enjoy their LED lighting. I plan to switch to the new LL140 models, which are a bit noisier but look great. If I weren’t so focused on my setup, I’d probably choose Noctua instead.

Setting up a custom curve was also problematic, and I hesitated to update Corsair Link for a long time. The latest version (after a clean install) seems to perform better now.

The main issue remains using AI Suite 3’s Fan Xpert for my case fans—this isn’t working well lately due to Windows 10 Update 10. This might be connected to the pump speed setting, which was at 100% before, and I have no idea what it’s currently set to.

I can adjust settings when I get home, but until then, I’m trying to determine if I can optimize my rig with what I have. I’m unsure yet, but I’m concerned the temperatures are already high, so I might not be able to push far.
M
MX_Flame
12-28-2017, 04:48 AM #11

I believe a tailored fan profile in Corsair Link is possible, though it's been quite some time since I last used it—it was quite unstable for me, but it's been over a year. The stock fans were removed and replaced with quieter Cougar models, which helped boost the RPM.

I can push the RPM significantly. My ML140 Pro fans run smoothly up to 56% speed, while the rest of my setup remains nearly silent. The only noticeable sound now is air movement. The CD drive, thanks to the ROG software I installed, has been causing issues all day—about four to five times louder than the rest of the rig.

I keep them mainly because Corsair makes excellent fans and I enjoy their LED lighting. I plan to switch to the new LL140 models, which are a bit noisier but look great. If I weren’t so focused on my setup, I’d probably choose Noctua instead.

Setting up a custom curve was also problematic, and I hesitated to update Corsair Link for a long time. The latest version (after a clean install) seems to perform better now.

The main issue remains using AI Suite 3’s Fan Xpert for my case fans—this isn’t working well lately due to Windows 10 Update 10. This might be connected to the pump speed setting, which was at 100% before, and I have no idea what it’s currently set to.

I can adjust settings when I get home, but until then, I’m trying to determine if I can optimize my rig with what I have. I’m unsure yet, but I’m concerned the temperatures are already high, so I might not be able to push far.

E
Eduardo_GameOn
Posting Freak
921
12-31-2017, 06:04 AM
#12
I believe the Rockit 88 kit is worth looking into, especially after performing the delid and using the AIO cooler. Once everything is set up, you shouldn't have to worry about temperature control, and the main constraint will be voltage. I've also been thinking about this myself, as I have some flexibility with my CPU's voltage and think I could push my 4770K to around 4.6-4.7.
E
Eduardo_GameOn
12-31-2017, 06:04 AM #12

I believe the Rockit 88 kit is worth looking into, especially after performing the delid and using the AIO cooler. Once everything is set up, you shouldn't have to worry about temperature control, and the main constraint will be voltage. I've also been thinking about this myself, as I have some flexibility with my CPU's voltage and think I could push my 4770K to around 4.6-4.7.

W
Woely
Member
108
01-03-2018, 08:58 AM
#13
I believe you should look into the Rockit 88 kit, especially after performing the delid and using an AIO cooler; then temperature shouldn't be a concern, and the main limitation would be voltage. I've been thinking about doing it too, since I have some extra space for my CPU voltage and think I could push my 4770K to around 4.6-4.7.
I suppose I need to. People seem to gain about 16°C from a delid of the 8700k, which is quite a jump.
I'll check what temperatures I can achieve with a modest overclock, and if it's acceptable I won't worry, but if not I'll explore that kit further.
W
Woely
01-03-2018, 08:58 AM #13

I believe you should look into the Rockit 88 kit, especially after performing the delid and using an AIO cooler; then temperature shouldn't be a concern, and the main limitation would be voltage. I've been thinking about doing it too, since I have some extra space for my CPU voltage and think I could push my 4770K to around 4.6-4.7.
I suppose I need to. People seem to gain about 16°C from a delid of the 8700k, which is quite a jump.
I'll check what temperatures I can achieve with a modest overclock, and if it's acceptable I won't worry, but if not I'll explore that kit further.

B
BBballoonboy
Junior Member
49
01-03-2018, 03:34 PM
#14
Because of concerns about the clock speeds and the high temperature, especially with water cooling, you might want to have Silicon lottery delid it for you. They’re affordable and offer a quick turnaround. Removing the chip can significantly reduce temperatures beyond what most people anticipate.
B
BBballoonboy
01-03-2018, 03:34 PM #14

Because of concerns about the clock speeds and the high temperature, especially with water cooling, you might want to have Silicon lottery delid it for you. They’re affordable and offer a quick turnaround. Removing the chip can significantly reduce temperatures beyond what most people anticipate.

F
fireroster
Member
150
01-03-2018, 03:53 PM
#15
Since you're concerned about the heat levels (and ~90°C not exceeding the clock speed is quite high, especially with water cooling), why not have Silicon's Delid it for you? They don't charge much and offer a one-day turnaround. Delidding could actually reduce temperatures more than most people expect. https://www.youtube.com/watch?v=HvgjbrmT3us I own an 8700k. I can delide it myself, but given the relatively low cost of having someone with extensive experience do it, I'm opting that way. I agree with the wildcard that the x62 is a better water cooler than the Corsair. Regarding the 90°C temperature, what temperatures can you expect in an Aida64? I found it to be a more accurate stress test than RealBench or Prime95. You might manage to lower it slightly (just a few degrees, in single digits) by using Kryonaut and being careful with the amount used. I don't think the x62 is significantly better than the x61, and the H115i (the top Corsair model I know of) outperforms the x61 by about 5-7°C. Even if the x62 beats the x61 by around 10°C (which seems reasonable), it would still only surpass the Corsair by a couple of degrees. I can't justify spending $160 on a slightly better cooler than my $140 one. The same applies to Kryonaut TG. I use an Arctic MX-4, and TG only cools 1.5-2°C more. It's not worth changing it unless my components are in good shape. It's not just about them doing it; I'd trust them more, but the turnaround time is still a big deal—even a day is much less than shipping and costs. It would cost me $30 to have them do it, then I'd have to ship it back, which would take another 2-3 days, and my PC would be down for a week. That's not something I'm willing to risk. I'd rather try it myself. I also came across this review: I really need to conduct more testing. It's possible I was focusing on the maximum temperature and saw it spike to around 83°C, which matches what you see here. My gaming loads run much cooler than these, so it's entirely possible the chip is just overheating as it is. The 90°C figure is likely a theory, not something I've verified in RealBench. Once I test it further, I'll know for sure. But at this point, it seems a delid is definitely necessary. Those temperatures are terrible. You'd expect Intel to have addressed this by now.
F
fireroster
01-03-2018, 03:53 PM #15

Since you're concerned about the heat levels (and ~90°C not exceeding the clock speed is quite high, especially with water cooling), why not have Silicon's Delid it for you? They don't charge much and offer a one-day turnaround. Delidding could actually reduce temperatures more than most people expect. https://www.youtube.com/watch?v=HvgjbrmT3us I own an 8700k. I can delide it myself, but given the relatively low cost of having someone with extensive experience do it, I'm opting that way. I agree with the wildcard that the x62 is a better water cooler than the Corsair. Regarding the 90°C temperature, what temperatures can you expect in an Aida64? I found it to be a more accurate stress test than RealBench or Prime95. You might manage to lower it slightly (just a few degrees, in single digits) by using Kryonaut and being careful with the amount used. I don't think the x62 is significantly better than the x61, and the H115i (the top Corsair model I know of) outperforms the x61 by about 5-7°C. Even if the x62 beats the x61 by around 10°C (which seems reasonable), it would still only surpass the Corsair by a couple of degrees. I can't justify spending $160 on a slightly better cooler than my $140 one. The same applies to Kryonaut TG. I use an Arctic MX-4, and TG only cools 1.5-2°C more. It's not worth changing it unless my components are in good shape. It's not just about them doing it; I'd trust them more, but the turnaround time is still a big deal—even a day is much less than shipping and costs. It would cost me $30 to have them do it, then I'd have to ship it back, which would take another 2-3 days, and my PC would be down for a week. That's not something I'm willing to risk. I'd rather try it myself. I also came across this review: I really need to conduct more testing. It's possible I was focusing on the maximum temperature and saw it spike to around 83°C, which matches what you see here. My gaming loads run much cooler than these, so it's entirely possible the chip is just overheating as it is. The 90°C figure is likely a theory, not something I've verified in RealBench. Once I test it further, I'll know for sure. But at this point, it seems a delid is definitely necessary. Those temperatures are terrible. You'd expect Intel to have addressed this by now.

D
davachio
Member
204
01-21-2018, 11:47 AM
#16
I've got my i8700k (air cooler, Bequiet Dark Rock Pro 3, not delidded) stable at 4.8Ghz with 1.25V. Gaming temperatures sit in the 50s, benchmarks and stress tests run from 60C to 70C, Prime95 V26.6 stays under 85C, and the latest Prime 95 version with AVX hits 90C (without any AVX offset). Overclocking without delidding is possible, but it performs better when using delidding.
D
davachio
01-21-2018, 11:47 AM #16

I've got my i8700k (air cooler, Bequiet Dark Rock Pro 3, not delidded) stable at 4.8Ghz with 1.25V. Gaming temperatures sit in the 50s, benchmarks and stress tests run from 60C to 70C, Prime95 V26.6 stays under 85C, and the latest Prime 95 version with AVX hits 90C (without any AVX offset). Overclocking without delidding is possible, but it performs better when using delidding.

O
Oxopvp80
Member
183
01-21-2018, 07:35 PM
#17
I have my i8700k (air cooler, Bequiet Dark Rock Pro 3, not delidded) stable overclocked to 4.8Ghz at 1.25V. Gaming temperatures are in the 50s, benchmarks and stress tests reach 60Cs to 70Cs, Prime95 V26.6 under 85C, Prime 95 latest version with AVX, hits 90Cs (without avx offset). Overclocking without delidding works, but it's better with delidding.
That seems like a really solid chip then.
😀
Congratulations! Lol.
I'm planning to do more testing. It might be that I'm only seeing spikes and not the average temperatures.
O
Oxopvp80
01-21-2018, 07:35 PM #17

I have my i8700k (air cooler, Bequiet Dark Rock Pro 3, not delidded) stable overclocked to 4.8Ghz at 1.25V. Gaming temperatures are in the 50s, benchmarks and stress tests reach 60Cs to 70Cs, Prime95 V26.6 under 85C, Prime 95 latest version with AVX, hits 90Cs (without avx offset). Overclocking without delidding works, but it's better with delidding.
That seems like a really solid chip then.
😀
Congratulations! Lol.
I'm planning to do more testing. It might be that I'm only seeing spikes and not the average temperatures.

E
Eragon1512
Member
94
01-21-2018, 07:51 PM
#18
EpIckFa1LJoN: bjornl: You're concerned about the temperatures (and ~90c not being overclocked is quite high, especially with water cooling). Why not have Silicon's Delid it for you? They don't charge much and offer a one-day turnaround. Delidding can actually reduce temperatures more than most people expect. https://www.youtube.com/watch?v=HvgjbrmT3us I own an 8700k. I can do the delidding myself, but given the cost of having someone with extensive experience handle it, I'm opting that way. I agree with the idea that the x62 is a better water cooler than the Corsair. Regarding the 90c temperature, what readings do you get in Aida64? I found it to be a much more accurate stress test than RealBench or Prime95. You might manage to lower it slightly (just a few degrees, in single digits) by using kryonaut and being careful with the amount. I don't think the x62 is significantly better than the x61, and the H115i (the top Corsair model I know of) outperforms the x61 by about 5-7C. Even if the x62 beats the x61 by around 10C, it would still lag behind the Corsair by just a couple degrees. I can't justify spending $160 on a slightly better cooler than my $140 one. The same applies to Kryonaut TG. I use an Arctic MX-4 and TG only cools 1.5-2C more. It's not worth changing it unless my current setup is failing completely. It's not just about them doing it; I'd trust them more, but the shipping time and cost are still a big deal. It's already $30 for their service, then I have to ship it back, which would take another 2-3 days, plus another week of downtime. My PC would be unusable for a week. I'd rather try it myself. I also found this review: I definitely need to test more. It's possible I was only checking the max temperature and saw it spike to about 83C, which matches what you're seeing here. My gaming loads are much cooler, so it's likely the chip is just running hot as it is. The 90C figure is probably a theory, not a real test. Once I try it out more, I'll know for sure. But right now, a delidding seems definitely necessary. Those temperatures are terrible. You'd think Intel would have fixed this by now. Good luck with your experiment. I'm sure it will work fine. I've done similar tests before using extreme cooling with Peltier junctions and a high-end water cooler. For now, I'm happy with a Noctua 15s. The pump won't fail, the parts won't need replacement. It's a simple, low-maintenance solution that matches your water cooler well. So we agree, good enough is fine. Regarding Kryonaut, you're right—it makes a small difference. Adding too much thermal compound (a common mistake) matters more than the type used. The price difference between them is negligible compared to the time and effort needed. It feels like running your CPU a bit hotter just because you're upset about it. I doubt it'll take a week in shipping to Texas, but I'll see. I don't mind much since I have other computers I can use for a few days.
E
Eragon1512
01-21-2018, 07:51 PM #18

EpIckFa1LJoN: bjornl: You're concerned about the temperatures (and ~90c not being overclocked is quite high, especially with water cooling). Why not have Silicon's Delid it for you? They don't charge much and offer a one-day turnaround. Delidding can actually reduce temperatures more than most people expect. https://www.youtube.com/watch?v=HvgjbrmT3us I own an 8700k. I can do the delidding myself, but given the cost of having someone with extensive experience handle it, I'm opting that way. I agree with the idea that the x62 is a better water cooler than the Corsair. Regarding the 90c temperature, what readings do you get in Aida64? I found it to be a much more accurate stress test than RealBench or Prime95. You might manage to lower it slightly (just a few degrees, in single digits) by using kryonaut and being careful with the amount. I don't think the x62 is significantly better than the x61, and the H115i (the top Corsair model I know of) outperforms the x61 by about 5-7C. Even if the x62 beats the x61 by around 10C, it would still lag behind the Corsair by just a couple degrees. I can't justify spending $160 on a slightly better cooler than my $140 one. The same applies to Kryonaut TG. I use an Arctic MX-4 and TG only cools 1.5-2C more. It's not worth changing it unless my current setup is failing completely. It's not just about them doing it; I'd trust them more, but the shipping time and cost are still a big deal. It's already $30 for their service, then I have to ship it back, which would take another 2-3 days, plus another week of downtime. My PC would be unusable for a week. I'd rather try it myself. I also found this review: I definitely need to test more. It's possible I was only checking the max temperature and saw it spike to about 83C, which matches what you're seeing here. My gaming loads are much cooler, so it's likely the chip is just running hot as it is. The 90C figure is probably a theory, not a real test. Once I try it out more, I'll know for sure. But right now, a delidding seems definitely necessary. Those temperatures are terrible. You'd think Intel would have fixed this by now. Good luck with your experiment. I'm sure it will work fine. I've done similar tests before using extreme cooling with Peltier junctions and a high-end water cooler. For now, I'm happy with a Noctua 15s. The pump won't fail, the parts won't need replacement. It's a simple, low-maintenance solution that matches your water cooler well. So we agree, good enough is fine. Regarding Kryonaut, you're right—it makes a small difference. Adding too much thermal compound (a common mistake) matters more than the type used. The price difference between them is negligible compared to the time and effort needed. It feels like running your CPU a bit hotter just because you're upset about it. I doubt it'll take a week in shipping to Texas, but I'll see. I don't mind much since I have other computers I can use for a few days.

R
RedCharizard
Member
65
01-21-2018, 10:16 PM
#19
You might be correct about the paste amount. But what impact would that have?
I used roughly a pea-sized dot in the center, about 1/4" diameter, just a bit wider than its height. It’s possible I underestimated the size of a pea. There could be a bit too much paste.
I didn’t use too much paste and think I should redo it.
You might be right—it may have been a little too long since my last application. I might try again when I get home to see if it improves things.
REVISIT: I’m not sure anymore... I saw a video for the paste I was using and they seem to use more than I did. Maybe I’ll just switch it up for safety xD.
R
RedCharizard
01-21-2018, 10:16 PM #19

You might be correct about the paste amount. But what impact would that have?
I used roughly a pea-sized dot in the center, about 1/4" diameter, just a bit wider than its height. It’s possible I underestimated the size of a pea. There could be a bit too much paste.
I didn’t use too much paste and think I should redo it.
You might be right—it may have been a little too long since my last application. I might try again when I get home to see if it improves things.
REVISIT: I’m not sure anymore... I saw a video for the paste I was using and they seem to use more than I did. Maybe I’ll just switch it up for safety xD.

D
dcchatter
Junior Member
6
01-25-2018, 05:34 PM
#20
EpIckFa1LJoN :
You might be correct about the paste. But what impact would that have?
I placed roughly a pea-sized dot in the center (about 1/4" diameter, just a bit wider than its height). I could have underestimated the size of a pea. It might be a bit too much paste.
I didn’t use too much paste. I think I should try again.
I usually apply less than a pea, closer to a grain of rice. The thermal compound is only meant to fill in tiny cracks and imperfections on both the cooler and CPU surfaces. From what you described, it seems like you might have applied a bit more paste. If my PC were running a little warm and I was concerned about the thermal compound, I’d just redo it.
D
dcchatter
01-25-2018, 05:34 PM #20

EpIckFa1LJoN :
You might be correct about the paste. But what impact would that have?
I placed roughly a pea-sized dot in the center (about 1/4" diameter, just a bit wider than its height). I could have underestimated the size of a pea. It might be a bit too much paste.
I didn’t use too much paste. I think I should try again.
I usually apply less than a pea, closer to a grain of rice. The thermal compound is only meant to fill in tiny cracks and imperfections on both the cooler and CPU surfaces. From what you described, it seems like you might have applied a bit more paste. If my PC were running a little warm and I was concerned about the thermal compound, I’d just redo it.

Pages (3): Previous 1 2 3 Next