X79 and/or rampage IV OC'ing thread for those of us still left....

[drizzler]

Hmm i am running 104,7 bclk without getting a "downgrade" to gen2 with my titan x maxwell on my rampage iv be. have you pcie spread spectrum enabled? could be the fury is more sensitive to pcie clock modulation compared to other cards, if so you should disable it. for pcie overclocking it should be disabled.

 

[Zyll Goliat]

Hmm i am running 104,7 bclk without getting a "downgrade" to gen2 with my titan x maxwell on my rampage iv be. have you pcie spread spectrum enabled? could be the fury is more sensitive to pcie clock modulation compared to other cards, if so you should disable it. for pcie overclocking it should be disabled.

Yes I keep PCIE spread spectrum disabled....to be precise 104.3 is working fine on BCLK and pcie gen3....104.4-104.9 is starts as PCIe gen3 but it's not stable it will crash win....above 105 works fine BUT downgrade on PCIe Gen2 automatically.....Honestly it's not a big deal I prefer that my CPU works on higher frequency (115.2bclk)and no PCIe gen3 as I can't see any difference in gaming/benchmarks at all....This was just issue that was driving me MAD but now that I figure what is all about I am totally fine with that....

 

[lightzout]

Just saw this thread. I have an Asus P9X79 LE that I would like to start pushing if possible to get better gaming performance playing Battlefield. My powersupply connector melted into the mobo at the top for the CPU but when I cleaned out themelted plastic from the PSU the mbo and the pin looked fine. Booted right up. I will review all the posts for BIOS settings. I found this on the side of the road in the woods with two other older PCs waiting to go to landfill. Its a cyperpow that shipped with dual/SLI 550tis. It took me a while to get it to boot up but once I updated to the latest bios and changed setting for windows 10 its been the best motherboard I have owwned at any price. I am willing to push if it helps otherwise I err on the side of stability. Any advice? I also have a R9 290 Vapor-x with stock settings which could be pushed if anyone has feedback there. That card has three fans. I feel obligated to reset thermal paste and see if I can push it further. Thoughts?Advice? Dad jokes?

 

[lexluthermiester]

Just saw this thread. I have an Asus P9X79 LE that I would like to start pushing if possible to get better gaming performance playing Battlefield. My powersupply connector melted into the mobo at the top for the CPU but when I cleaned out themelted plastic from the PSU the mbo and the pin looked fine. Booted right up. I will review all the posts for BIOS settings. I found this on the side of the road in the woods with two other older PCs waiting to go to landfill. Its a cyperpow that shipped with dual/SLI 550tis. It took me a while to get it to boot up but once I updated to the latest bios and changed setting for windows 10 its been the best motherboard I have owwned at any price. I am willing to push if it helps otherwise I err on the side of stability. Any advice? I also have a R9 290 Vapor-x with stock settings which could be pushed if anyone has feedback there. That card has three fans. I feel obligated to reset thermal paste and see if I can push it further. Thoughts?Advice? Dad jokes?

What CPU/RAM do you have? This will help us give you better advice.

And welcome to the thread!

 

[lightzout]

Sorry I thought I had included the details. I have a P9X79 LE with Intel(R) Core(TM) i7-3820 CPU @ 3.60GHz, 18 GB RAM, evo ssd, R9 290 Vapor-x TRI OC 4GB. It is an ancient PC built for Windows 8 and it has flaws but it is surprisingly robust. The PSU cable to the CPU melted but I got it out and cleaned up, installed a new PSU and it booted right up. I would like to OC a little and/or bios flash GPU.

 

[drizzler]

Are you willing to spent a little amount to improve the performance? I would switch the 3820 with at least an e5 1650v2 (6cores) that can hit 4,4+ghz easily most of the time. You can get this cpu at around 50 Euros. The 1680v2 (8cores) would be the ultimate overclockable solution but the more expensive one.

What is your cooling solution?

 

[Zyll Goliat]

Sorry I thought I had included the details. I have a P9X79 LE with Intel(R) Core(TM) i7-3820 CPU @ 3.60GHz, 18 GB RAM, evo ssd, R9 290 Vapor-x TRI OC 4GB. It is an ancient PC built for Windows 8 and it has flaws but it is surprisingly robust. The PSU cable to the CPU melted but I got it out and cleaned up, installed a new PSU and it booted right up. I would like to OC a little and/or bios flash GPU.

There are plenty videos and guides online how to OC 3820...Just tried that it's going to be easy if you have at least some experience...nothing to be afraid if it doesn't post after bad OC once or twice it will load default clock speeds and you will be able to get into the bios again....after you reach the windows you should do some benchmarks and compare your results(OC Vs Non OC) but also pay attention on your heat and CPU V and then see do you like it or not....and yes you can share all of that with us and ask any questions.....3820 Non K is a bit more difficult to OC but also it's not that hard at all,I "played" before with one of those and it was super-easy.....

Here this is just one of the guide that seems OK LINK

 

[lexluthermiester]

I have a P9X79 LE with Intel(R) Core(TM) i7-3820 CPU @ 3.60GHz,

IMO, that 3820 is past it's prime and long in the tooth for gaming, even at 1080p. An upgrade to a 6 or 8 core CPU and overclocking it would give you a very solid performance boost. You board supports Xeons as long as it's up to date:
https://www.asus.com/us/SupportOnly/P9X79 LE/HelpDesk_CPU/

@drizzler's suggestion of a Xeon E5-1650V2 is a great option. A Xeon E5-1680V2 or a Xeon E5-2667V2 are also good options. Example below;

Intel Xeon E5-1650 v2 3.5GHz LGA2011 6-Cores 12MB 130W CPU Processor SR1AQ | eBay

Baltimore, Maryland, 21230. Item Location :-.

www.ebay.com

1650V2 $50 shipped

Intel Xeon E5-1680V2 SR1MJ 3.00 Ghz 8-Core Processor OEM | eBay

Working pulls. iPower Resale is a small, passionate group specializing in the resale of open box, refurbished, and pre-owned Apple hardware. Condition of each product may. and type of product you are buying.

www.ebay.com

1680V2 $130 shipped

SR19W INTEL XEON 8 CORE CPU E5-2667V2 25M CACHE 3.30 GHZ CPU CM8063501287304 | eBay

Find many great new & used options and get the best deals for SR19W INTEL XEON 8 CORE CPU E5-2667V2 25M CACHE 3.30 GHZ CPU CM8063501287304 at the best online prices at eBay! Free shipping for many products!

www.ebay.com

2667V2 $140 shipped

The 1650v2 is the best deal and they are easily OC'd. The 8cores are still worth it if a bit pricier.

18 GB RAM

How is that set up? What are your DIMM sizes and how many are there?

 

[Zyll Goliat]

IMO, that 3820 is past it's prime and long in the tooth for gaming, even at 1080p. An upgrade to a 6 or 8 core CPU and overclocking it would give you a very solid performance boost. You board supports Xeons as long as it's up to date:
https://www.asus.com/us/SupportOnly/P9X79 LE/HelpDesk_CPU/

@drizzler's suggestion of a Xeon E5-1650V2 is a great option. A Xeon E5-1680V2 or a Xeon E5-2667V2 are also good options. Example below;

Intel Xeon E5-1650 v2 3.5GHz LGA2011 6-Cores 12MB 130W CPU Processor SR1AQ | eBay

Baltimore, Maryland, 21230. Item Location :-.

www.ebay.com

1650V2 $50 shipped

Intel Xeon E5-1680V2 SR1MJ 3.00 Ghz 8-Core Processor OEM | eBay

Working pulls. iPower Resale is a small, passionate group specializing in the resale of open box, refurbished, and pre-owned Apple hardware. Condition of each product may. and type of product you are buying.

www.ebay.com

1680V2 $130 shipped

SR19W INTEL XEON 8 CORE CPU E5-2667V2 25M CACHE 3.30 GHZ CPU CM8063501287304 | eBay

Find many great new & used options and get the best deals for SR19W INTEL XEON 8 CORE CPU E5-2667V2 25M CACHE 3.30 GHZ CPU CM8063501287304 at the best online prices at eBay! Free shipping for many products!

www.ebay.com

2667V2 $140 shipped

The 1650v2 is the best deal and they are easily OC'd. The 8cores are still worth it if a bit pricier.


How is that set up? What are your DIMM sizes and how many are there?

Sure 3820 past it's prime but then again properly OC it' going to be good enough to push that R9 290....many of those I7 3820 when working on 4.2-4,5Ghz still have more than a solid IPC problem is of course number of cores/threads but i am sure he will be ok with that setup if his plan is to game on 1080p/60hz/medium.........

 

[lexluthermiester]

Sure 3820 past it's prime but then again properly OC it' going to be good enough to push that R9 290.

Good point.

 

[Zyll Goliat]

Ok guys I did further testing regarding my PCIe GEN3 Vs PCIe GEN2 OC/BCLK "issue"...and now finally I have all info about it...So OK we now know that the cause for downgrading from PCIe 3 to the PCIe 2 was actually the BCLK OC as I already mention above so in short as we rising the bclk also PCIe clock getting higher but beside that I was really curious what are the actual speed on PCIe when I oveclock BCLK and I have some very interesting results.....I managed to have fully stable PCIe Gen3 to work on 104,2 BCLK and actual Interface Transfer Bandwith(video memory) was 8.58GB/s which ain't that badon normal/regular PCIe GEN 3 speed are around 8GB/s but when I OC my CPU on my fully stable maximum 115,2 BCLK then ofcourse PCIe GEN2 is only operational BUT keep in mind that now is highly overclocked so now the Interface Transfer Bandwith was 6,45GB/s which is waay more then regular/non OC PCIe GEN2 speed around 5GB/s)....

Those results are from SiSoftSandra benchmarks video memory transfer bandwith.........
P.S.In games and GPU Benchmarks the GPU results are more or less identical in fact they are a bit better with OC/PCIe GEN2 but that is also logical as my CPU then working on much higher clocks...
P.S.S.Those results are done with R9 Fury I am curious to know IF will be any difference with some much more powerful GPU that actually maybe can use that faster bandwith....

 

[drizzler]

PCIe 2.0 x16 bandwith is still enough bandwith for a rtx 3090, you need round about 2gb/s to render in 4k with 32 bit and 60fps so this is not a problem in the gpu department. PCIe 2.0 x16 can do 8GB/s

With NVME SSDs it is a different story cause they are only linked with x4 (each lane can do 500 mb/s), so with pcie 2.0 you are limited to round about 1.8gb/s (2gb/s - overhead) without overclocking the pcie frequency. So even an older Samsung 950 pro is limited with PCIE 2.0 at stock pcie frequency. Direct IO storage is something to consider too in this regard.

But one more question : with the Asus Sabertooth X79 do you have the tweakers paradise options to play arround with skews for cpu i/o etc? I do not have the facts for error handling in mind regarding pcie 3.0 but maybe there is a fall back routine if too many pcie errors are occuring and the link is falling back to 2.0. Maybe this is something the fury does off its own. If so adjusting the clock skew for pcie might fix this not so important problem.

 

[Zyll Goliat]

PCIe 2.0 x16 bandwith is still enough bandwith for a rtx 3090, you need round about 2gb/s to render in 4k with 32 bit and 60fps so this is not a problem in the gpu department. PCIe 2.0 x16 can do 8GB/s

With NVME SSDs it is a different story cause they are only linked with x4 (each lane can do 500 mb/s), so with pcie 2.0 you are limited to round about 1.8gb/s (2gb/s - overhead) without overclocking the pcie frequency. So even an older Samsung 950 pro is limited with PCIE 2.0 at stock pcie frequency. Direct IO storage is something to consider too in this regard.

But one more question : with the Asus Sabertooth X79 do you have the tweakers paradise options to play arround with skews for cpu i/o etc? I do not have the facts for error handling in mind regarding pcie 3.0 but maybe there is a fall back routine if too many pcie errors are occuring and the link is falling back to 2.0. Maybe this is something the fury does off its own. If so adjusting the clock skew for pcie might fix this not so important problem.

Ohh sure I know that PCIe 2 is enough even for 3090 but in that case you could potentially start loosing few frames here&there in some games/benchmarks while with "weaker" GPU like the Fury there is literally no difference at all....no there are no errors at all(in between 104,3-104,9 the win will just froze on higher bclk it will start with downgrading PCIe and everything will work totally normal) I believe that actual problem is probably HIGH PCIe link speed as after certain point of BCLK/OC this motherboard simply can not handle that speed anymore so it will downgraded on gen 2 but as I said and show on pic above that "downgraded" PCIe Gen 2 speed is now when is OC actually closer to the PCIe Gen3 speed then to the normal/regular PCIe Gen2 speed....Not sure but maybe you are correct and some other GPU could handle more BCLK/OC tho' I still believe it's more due to the limitation of the X79 motherboard....
Yeah this board is great for OC have also option for strap BUT it's like 100,125,150 unfortunately those Xeons can not handle strap on 125 or more like some unlocked I7.......

 

[Eil1999]

Hi everyone. Better late than never they say. Got myself p9x79 and 1680v2. Looking for some guidance with achieving balanced overclock (good for 24/7). After so many years Internet is full of OC guides but not many of exactly this chip. After reading this thread I know that some of you tested this chip in every way.

So, if there is guide that from your perspective works pretty well with this cpu I would appreciate your answer.
Also any tips are welcome.

 

[dalekdukesboy]

Hi everyone. Better late than never they say. Got myself p9x79 and 1680v2. Looking for some guidance with achieving balanced overclock (good for 24/7). After so many years Internet is full of OC guides but not many of exactly this chip. After reading this thread I know that some of you tested this chip in every way.

So, if there is guide that from your perspective works pretty well with this cpu I would appreciate your answer.
Also any tips are welcome.

Welcome! And abused this chip in every way is probably more accurate...

 

[Eil1999]

Welcome! And abused this chip in every way is probably more accurate...

I would say discovering it's max potential. Well, had a little bit of playing with it.
Made first part of testing with "raw" bios. Everything on Asus's defaults, only bclk set to 100 (on auto it likes to jump) and played with manual voltage and multiplier.
Results with L - passed 10 runs of linx, with C - passed 10 minutes of cinebench r23

4.2Ghz 1.15V C
4.2Ghz 1.19V L
4.3Ghz 1.2V C
4.3Ghz 1.23V L
4.4Ghz 1.25V C
4.4Ghz up to 1.35V L error

It seems that I'm hitting some power celling in this one. More probable that vrms or smth is getting hotter than surface of the sun. In this setup I'm running without any additional cooling and with an aio. This mobo is the cheapest asus model so that's may be it. After this I continued with cinebech tests only.

4.5Ghz 1.32V C
4.6Ghz up to 1.39V C error
Also tried if it launches at 4.7Ghz but 1.42V is not enough to make it to desktop.

Before the second series of tests I'm gonna make few changes. Gonna check with memtest if rams are good because I feel that something may be fishy with those. Gonna reapply thermal paste because I'm getting big differences between cores (up to 12 degrees). Will strap fans in key places and reinstall windows.

 

[dalekdukesboy]

I would say discovering it's max potential. Well, had a little bit of playing with it.
Made first part of testing with "raw" bios. Everything on Asus's defaults, only bclk set to 100 (on auto it likes to jump) and played with manual voltage and multiplier.
Results with L - passed 10 runs of linx, with C - passed 10 minutes of cinebench r23

4.2Ghz 1.15V C
4.2Ghz 1.19V L
4.3Ghz 1.2V C
4.3Ghz 1.23V L
4.4Ghz 1.25V C
4.4Ghz up to 1.35V L error

It seems that I'm hitting some power celling in this one. More probable that vrms or smth is getting hotter than surface of the sun. In this setup I'm running without any additional cooling and with an aio. This mobo is the cheapest asus model so that's may be it. After this I continued with cinebech tests only.

4.5Ghz 1.32V C
4.6Ghz up to 1.39V C error
Also tried if it launches at 4.7Ghz but 1.42V is not enough to make it to desktop.

Before the second series of tests I'm gonna make few changes. Gonna check with memtest if rams are good because I feel that something may be fishy with those. Gonna reapply thermal paste because I'm getting big differences between cores (up to 12 degrees). Will strap fans in key places and reinstall windows.

Interesting results....I just saw the core "differences" and minus just bad sensors they aren't usually that far apart even though they do separate a bit when under stress and there's a range...however on second thought that isn't much different than my "hottest" core versus my "coolest" one according to sensor. What kind of cooling are you using? That has a lot to do with results I have a good idea of what you can do without even knowing actual temps by cooling capacity.

I also would say YES to fans on vrm's etc because at various times I've seen them heat up quite a bit under stress and an overclock. I just saw you are using an (AIO) cooler upon rereading and taking head out of my ass reading it closer but which one? Even that could make a fair amount of difference! Unless you have a meh overclocking chip I'd be surprised if you can't take that quite a bit higher than where it fails with errors...

I had a chip I had degrade or stop working for whatever reason but it wasn't great OC and was used to start and got to 4.6-4.7 stable'ish under stress testing till it starting going down the tubes and eventually wouldn't boot at ANY speed including stock. That was my first cpu ever that pulled that stunt! This one I got is a fairly unique stepping and I'm guessing a better/latter one than most and it does indeed clock higher with significantly lower voltage than most chips I've seen including first one I had.

Anxious to see your results when you put some fans on key areas and reset paste etc!

 

[Eil1999]

Interesting results....I just saw the core "differences" and minus just bad sensors they aren't usually that far apart even though they do separate a bit when under stress and there's a range...however on second thought that isn't much different than my "hottest" core versus my "coolest" one according to sensor. What kind of cooling are you using? That has a lot to do with results I have a good idea of what you can do without even knowing actual temps by cooling capacity.

I also would say YES to fans on vrm's etc because at various times I've seen them heat up quite a bit under stress and an overclock. I just saw you are using an (AIO) cooler upon rereading and taking head out of my ass reading it closer but which one? Even that could make a fair amount of difference! Unless you have a meh overclocking chip I'd be surprised if you can't take that quite a bit higher than where it fails with errors...

I had a chip I had degrade or stop working for whatever reason but it wasn't great OC and was used to start and got to 4.6-4.7 stable'ish under stress testing till it starting going down the tubes and eventually wouldn't boot at ANY speed including stock. That was my first cpu ever that pulled that stunt! This one I got is a fairly unique stepping and I'm guessing a better/latter one than most and it does indeed clock higher with significantly lower voltage than most chips I've seen including first one I had.

Anxious to see your results when you put some fans on key areas and reset paste etc!

I'm running Arctic Liquid Freezer II 280. And with those 12 degrees difference it was during linx at 1.35V. In cinebench it was around 8-9.
And don't get me wrong, this is only for testing. I'm not gonna run this setup at highest possible clocks because I just don't need it. Gonna settle at some clocks around 1.2V for daily use.
Soo yeah, not extreme overclocking.
Funny thing, I have watt meter and checked how much it's pulling from the wall. Idle I'm getting around 130W and found that during stress tests it's getting errors (aka getting less stable) if it's pulling over 400W in cpu related work. So either some thermal related problems or maybe insufficient psu. I'm running 650W gold Seasonic from previous built and even tho on paper it should work, just maybe some nanosecond spikes might be a problem. Who knows.

 

[dalekdukesboy]

I'm running Arctic Liquid Freezer II 280. And with those 12 degrees difference it was during linx at 1.35V. In cinebench it was around 8-9.
And don't get me wrong, this is only for testing. I'm not gonna run this setup at highest possible clocks because I just don't need it. Gonna settle at some clocks around 1.2V for daily use.
Soo yeah, not extreme overclocking.
Funny thing, I have watt meter and checked how much it's pulling from the wall. Idle I'm getting around 130W and found that during stress tests it's getting errors (aka getting less stable) if it's pulling over 400W in cpu related work. So either some thermal related problems or maybe insufficient psu. I'm running 650W gold Seasonic from previous built and even tho on paper it should work, just maybe some nanosecond spikes might be a problem. Who knows.

Ok I've heard of cooler but will look up to see where it is in the "standings" so to speak especially if you are using the stock fans that come with it. Watt meters are neat, just sayin'. I admit lack of pushing chip to limit "extreme oc'ing" as you called it makes me sad, but that's just me and I realize you aren't looking to keep it 24/7 at really high clocks which is all good I am glad seeing you are pushing it to see what it's upper limits are . Anyway what kind of temps are you getting at various clocks and voltages? I'd be very interested to see some screenshots of your results particularly if you can show us temps for all cores and linx version you're running and so forth.

 

[Eil1999]

Ok I've heard of cooler but will look up to see where it is in the "standings" so to speak especially if you are using the stock fans that come with it. Watt meters are neat, just sayin'. I admit lack of pushing chip to limit "extreme oc'ing" as you called it makes me sad, but that's just me and I realize you aren't looking to keep it 24/7 at really high clocks which is all good I am glad seeing you are pushing it to see what it's upper limits are . Anyway what kind of temps are you getting at various clocks and voltages? I'd be very interested to see some screenshots of your results particularly if you can show us temps for all cores and linx version you're running and so forth.

Yeah, I'm using stock fans. Okey, replaced thermal paste (still same difference between cores), put some old fan to blow just a little over vrms and decided to manually set fans to ramp up a little bit sooner. Don't have time to run all those linxs now but it seams that errors are still where they were.
But not all bad. I've managed to pass 10 minutes of r23 at 4,6Ghz (1.4V and if we believe cpuz undervolts to 1.368V during load). Max temps:

Also max temps at 4.2 with r23 (1.15V, undervolts to 1.128V):

And from nowhere actually booted at 4.7Ghz (1.42V) which is weird because before it didn't wanted to boot at that voltage.

But have in mind that I didn't touch anything besides clock multiplier and vcore voltage in bios. Rest is set to default so probably there is some "protection" that kills my stability for some reason.

And with linx I'm running 0.9.11, 10 passes with 25600MB of memory used.

 

[dalekdukesboy]

Yeah, I'm using stock fans. Okey, replaced thermal paste (still same difference between cores), put some old fan to blow just a little over vrms and decided to manually set fans to ramp up a little bit sooner. Don't have time to run all those linxs now but it seams that errors are still where they were.
But not all bad. I've managed to pass 10 minutes of r23 at 4,6Ghz (1.4V and if we believe cpuz undervolts to 1.368V during load). Max temps:
View attachment 202623
Also max temps at 4.2 with r23 (1.15V, undervolts to 1.128V):
View attachment 202625
And from nowhere actually booted at 4.7Ghz (1.42V) which is weird because before it didn't wanted to boot at that voltage.
View attachment 202624

But have in mind that I didn't touch anything besides clock multiplier and vcore voltage in bios. Rest is set to default so probably there is some "protection" that kills my stability for some reason.

And with linx I'm running 0.9.11, 10 passes with 25600MB of memory used.

I'm not surprised it's behaving better if you have fans over vrms. Temps are really low but I've never used r23 and I'm assuming Linx stresses it far harder, I know for me it gets highest temps of anything I've ever run.

 

[drizzler]

Cinebench is far way from beeing a viable stability stress test utitlity, no matter what version. for doing some bench runs or even quick checking voltage or memory settings it is ok.

Even IBT / LinX or Prime95 on its own are not the best option to test 24/7 realworld usage stability cause of its constant load. You will slightly not test your voltage / vrm settings / transient response with it cause of the missing variable load. You either have to move your mouse constantly (to trigger windows timer events that will cause load changes) or play a video in parallel to test.

OCCT is a better tool in this regard with its variable load but from my experience there are some scenarions in which you can pass OCCT while failing the other ones, looks like the avx usage is not that heavy with OCCT compared to the mentioned other ones.

 

[Eil1999]

Cinebench is far way from beeing a viable stability stress test utitlity, no matter what version. for doing some bench runs or even quick checking voltage or memory settings it is ok.

Even IBT / LinX or Prime95 on its own are not the best option to test 24/7 realworld usage stability cause of its constant load. You will slightly not test your voltage / vrm settings / transient response with it cause of the missing variable load. You either have to move your mouse constantly (to trigger windows timer events that will cause load changes) or play a video in parallel to test.

OCCT is a better tool in this regard with its variable load but from my experience there are some scenarions in which you can pass OCCT while failing the other ones, looks like the avx usage is not that heavy with OCCT compared to the mentioned other ones.

I wasn't using cinebench as a viable stress test but only to find voltage where pc will not bsod from opening single app or smth. I know it's far from stable.
Also I agree with linx. For me it's similar to prime95 on small ffts. It's just not reasonable realworld workload. Every person has different opinion on what to use to check stability so you never can be sure that your system is definitely stable.

 

[Zyll Goliat]

Cinebench is far way from beeing a viable stability stress test utitlity, no matter what version. for doing some bench runs or even quick checking voltage or memory settings it is ok.

Even IBT / LinX or Prime95 on its own are not the best option to test 24/7 realworld usage stability cause of its constant load. You will slightly not test your voltage / vrm settings / transient response with it cause of the missing variable load. You either have to move your mouse constantly (to trigger windows timer events that will cause load changes) or play a video in parallel to test.

OCCT is a better tool in this regard with its variable load but from my experience there are some scenarions in which you can pass OCCT while failing the other ones, looks like the avx usage is not that heavy with OCCT compared to the mentioned other ones.

Actually there is a huge difference between Cinebench R15 and Cinebench R23.....CineBench R23 is so much more demanding test and beside that you need much more time to do complete test and it's not bad at all to check your CPU stability......

 

[pevac]

In these 20 years of overclocking, no program is 100 percent secure.
I release prime 95, after 2 hours I release IBT 100 lap without shutting down prime, if the test doesn't break I finally lower 50-100mhz and that's it.