The Truth About CPU Soldering

[las]

What pump are you using? Corsair AIO pumps are hardly considered silent?!

You just sound mad, that's all

 

[EarthDog]

What pump am I using on an AIO? LOL, the one that comes with it obviously. Again, it makes noise as does every other setup, and it ramps up by default like every other setup. But when I am running games for review, its no louder than anything else sitting in this room (which was also made for quiet, hence the Yate Loon Fans and low RPM on the daily driver). The same AIO will ramp up proportionately with an AMD CPU as well, don't forget. They run by temperature, not by power use.

 

[las]

What pump am I using on an AIO? LOL, the one that comes with it obviously

Yeah, then it's not what I would consider silent. Far from it.

Yep they run by temp and Intel cpu's run hotter which means more RPM on pump. Cheap low-end pump struggling = Noisy.

You can replace fans all you want, but the loudest part of those cheap AIO's are the pump.

Yate Loon fans are pretty much as low end as it gets

I prefer my D5 vario

 

[EarthDog]

This back and forth is useless. Clearly your intent here is to be declarative and nothing will stop you from getting what you want to say out, even if it based on nothing. Enjoy your Intel CPU while you are reaching around juggling AMD in the other hand.

Cheers.

 

[qubit]

Just pause it there for a minute lads, I'm off to get another giant bucket of popcorn!

 

[EarthDog]

I'm done... this kid is off his rocker.

 

[dorsetknob]

@las
Over several threads you seem to be trying to "pick a fight "
suggest you dial it back or try those tactics with a few of our Mods
I will Say it now before its to late
"Thank you for Participating here on the Forums"

 

[qubit]

I think it's safe to say that the bottom line is that the Intel CPUs have the better performance and it doesn't matter how they did it. You want the best performing CPU? Go Intel. Could they work even better? Sure, but the beancounters there did their sums (cost-benefit analysis) and figured out that there was still enough performance left with the current cheaper thermal paste that it wasn't worth spending more to get that better performance. Not great for us enthusiasts, but that's how it is.

In my opinion, had AMD come out with a CPU that used a soldered die, had a better IPC and better clock rates, more cores and obliterated Intel in benchmarks, then you can bet your bottom dollar that Intel would have eeked every last ounce of performance out of their CPUs with better paste etc and likely updated their architecture to improve IPC too. But alas, that's not where we are today. It looks like Ryzen v2 will go some way to address this. We can only wait (impatiently).

 

[RCoon]

I prefer my D5 vario

Lowest noise level on a D5 Vario I've seen is 31 dB, and that's decoupled from the case at a very low flow rate.
Your system specs say ~20.

Yate Loon fans are pretty much as low end as it gets

mkay we'll take your opinion as gospel despite three reviews I have to hand that say otherwise.

Why are you mad?

This kills the crab.

In short: When will the world start posting information with citations. Tech forums require science-based evidence as opposed to "lol your fans suck. y u mad bro" - which has absolutely nothing to do with solder vs paste I might add.

 

[Space Lynx]

hey guys I just finished my build with an 8700k last night, 3 hours of 5ghz 1.365v -2 avx and no crashes, and temps never went past 76 celsius. no delid, but i did put conductonaut on the cpu and noctua heatsink. (i was very safe, not my first rodeo with the stuff).

my only problem is, i downclocked to 4800 mhz the entire time it was on prime95, as soon as i end prime95 it goes back up to umm 5ghz even the entire time, what exactly am I doing wrong, and how come the vcore is not staying at 1.365v, but only maxing out at 1.34v? man i suck at OC'ing... i used to be so good on sandy bridge, all this new stuff though... ugh

 

[EarthDog]

Should have started your own thread...

Its likely the AVX offset of -2 you set. Check bios and set to zero. If that doesnt do it, try running intel xtu and look at the bottom for throttle reasons (power limit, thermal, vrm throttling), you may need to resize the area its in to see it all.

Re: vcore...probably vdrop and vdroop. LLC combats this... just know now you are stable at a load of 1.34v.

 

[Space Lynx]

Should have started your own thread...

Its likely the AVX offset of -2 you set. Check bios and set to zero. If that doesnt do it, try running intel xtu and look at the bottom for throttle reasons (power limit, thermal, vrm throttling), you may need to resize the area its in to see it all.

Re: vcore...probably vdrop and vdroop. LLC combats this... just know now you are stable at a load of 1.34v.

Thank you senpai, I will tinker for a few more days and a make a topic if I don't get it hammered down by then.

 

[evernessince]

I believe that my CPU's above average performance is due to a lucky minimum gap. I have the same TIM as everyone else. And my VID is nothing special. I am also using an air cooler and arctic silver 5, so nothing exotic to cool it either.

As far as the theory relating to the gap, that is a pretty popular idea. You can use google and find people claiming that the chief benefit of moving from the original paste + RTV glue to delidding and using gallium on the die and superglue to reattach the lid is that the gap is narrower so the lid is closer to the die.

If you ever repaste a laptop, you will find that you need to use a copper shim to replace the original thermal pad to close the gap between the die and the heatsink. Running a thick layer of thermal paste does not work very well, you need a shim to transfer. It is a similar idea.

The problem is the black RTV adhesive that Intel uses to adhere the lid to the circuit board. It is thick, possibly to reduce damage to the die from overclamping. The thick glue causes the lid to be further from the die, so the thermal paste is excessively thick or even gapped and does not perform well at transferring between the die and the lid.


There isn't any. Just imaginary E-Peen.

Thermal pads are thicker than thermal paste by nature. Using them as an example for desktop CPUs is Apples to Oranges, it simply isn't the same thing. Don't think random people on forums counts as anything remotely considered evidence.

If it were really as simple as using super glue, don't you think Intel would have done that when a bunch of randos on the internet can?

I think it's safe to say that the bottom line is that the Intel CPUs have the better performance and it doesn't matter how they did it. You want the best performing CPU? Go Intel. Could they work even better? Sure, but the beancounters there did their sums (cost-benefit analysis) and figured out that there was still enough performance left with the current cheaper thermal paste that it wasn't worth spending more to get that better performance. Not great for us enthusiasts, but that's how it is.

In my opinion, had AMD come out with a CPU that used a soldered die, had a better IPC and better clock rates, more cores and obliterated Intel in benchmarks, then you can bet your bottom dollar that Intel would have eeked every last ounce of performance out of their CPUs with better paste etc and likely updated their architecture to improve IPC too. But alas, that's not where we are today. It looks like Ryzen v2 will go some way to address this. We can only wait (impatiently).

This is pretty questionable, given that AMD's SMT implementation is superior and that they can pack more cores on a single CPU thanks to their infinity fabric. Intel CPUs win in a single metric, single threaded performance, and loose in every other one. Power Consumption, Price, Temps, Core density, ect. AMD's top end CPU sells for half the price of Intel's and packs 6 more cores and better multi-threaded performance. Your comment is just that opinion devoid of fact.

Now get back on topic, because this thread has nothing to do with AMD vs Intel.

 

[delshay]

Yeah, we know the paste can be better, but its clearly good enough to overclock their chips quite a bit. Im sorry you are bitter a couple hundred mhz were left on the table. Even with that paste, the intel cpu overclocks more. AMD is whose clearly tapped out currently. Their cpus can get past their own xfr and like 200 mhz past their own all core boost. Intel can with their 'crappy' paste.

I dont care about power efficiency too much. Id rather have a 95w chip at 5 ghz (or 4.8 in your world) than a 65w chip locked at 4.1. That refresh is really pushing the envelope gaining an extra 100-200 mhz, lol! I bet they wont be able to overclock much (100-200 mhz) past their xfr until Zen2 is released in 2019.

Amd still has some work to do too. Competition is nice, but make no mistake about it, in some areas, including IPC and overclocking, they have some work to do.

id start a thread of your own.

I only posted that information for the user that got bored of seeing the same boring thing over & over again. Now I am going to post another idea.

How about side contact with the IHS. ie you have a normal heatsink that have four side bars built into the heatsink. The idea here is you mount the heatsink as normal, then you use a allen key to screw tighten two of the four side bars to clamp the IHS. The idea here is you not only got contact from the top of the IHS, but also from the side. The heatsink now has a bigger contact patch with the IHS which should transfer more heat to the heatsink. In other words you are squeezing the IHS from all four sides.

I have other strange/weird modifications here, some are fully working, but I have not posted much information on them & some have never being done before, it's a first. Some of these modifications are on my main classic computer, as it works so well.

 

[dorsetknob]

SORRY BUT THIS IDEA IS SO FUNNY AND A QUICK WAY OF BORKING A MODERN CPU

How about side contact with the IHS. ie you have a normal heatsink that have four side bars built into the heatsink. The idea here is you mount the heatsink as normal, then you use a allen key to screw tighten two of the four side bar to clamp the IHS. The idea here is you not only got contact from the top of the IHS, but also from the side.

All you would do is BELL the IHS and Break the TIM Bond Resulting in a Runaway cooking CPU

PS if you want to keep posting these ideas CREATE YOUR OWN THREAD thank you

 

[delshay]

SORRY BUT THIS IDEA IS SO FUNNY AND A QUICK WAY OF BORKING A MODERN CPU



All you would do is BELL the IHS and Break the TIM Bond Resulting in a Runaway cooking CPU

PS if you want to keep posting these ideas CREATE YOUR OWN THREAD thank you

You do not overtighten it. Have you seen AMD IHS delidded, you are not going to bend that so easy unless you go crazy on it.

EDIT: Let me see if I can grab a screenshot of a AMD IHS delidded, it's pretty thick not easy to bend or bow.

EDIT2: Added AMD IHS

 

[qubit]

Now get back on topic, because this thread has nothing to do with AMD vs Intel.

Cheers for telling me what to do, bud.

I guess this gives me a second reason for not bothering to reply to the rest of your post other than to say that you're coming off as another AMD apologist.

 

[dyonoctis]

I only posted that information for the user that got bored of seeing the same boring thing over & over again. Now I am going to post another idea.

How about side contact with the IHS. ie you have a normal heatsink that have four side bars built into the heatsink. The idea here is you mount the heatsink as normal, then you use a allen key to screw tighten two of the four side bars to clamp the IHS. The idea here is you not only got contact from the top of the IHS, but also from the side. The heatsink now has a bigger contact patch with the IHS which should transfer more heat to the heatsink. In other words you are squeezing the IHS from all four sides.

I have other strange/weird modifications here, some are fully working, but I have not posted much information on them & some have never being done before, it's a first. Some of these modifications are on my main classic computer, as it works so well.

To me that sound overly complicated for a gain that would be minimal...

 

[delshay]

To me that sound overly complicated for a gain that would be minimal...

It's not complicated. It will all be pre-attached to the heatsink. You mount heatsink as normal, then tighten two additional screws for clamping.

But if this all sounds too crazy I will delete posting.

 

[DRDNA]

This experiment

watch this space for improvements.

Note: Not all experiments work here & some do fail.

I love it that you are taking time to see if there is room for improvement in this area!

To me that sound overly complicated for a gain that would be minimal...

Well this guy seems like a real TPU-er from the old days where we ALWAYS were looking for any edge to help the overclock or the bench or the simple FPS.....glad to see still some folks like this left on TPU. delshay

 

[delshay]

I love it that you are taking time to see if there is room for improvement in this area!

Well this guy seems like a real TPU-er from the old days where we ALWAYS were looking for any edge to help the overclock or the bench or the simple FPS.....glad to see still some folks like this left on TPU. delshay

I like to experiment with different ways of doing things. Not the same thing over & over. I hardware mod on an old computer & if it works I then transfer it to my main computer. So my main computer is never at risk.

 

[EarthDog]

Kudos and all, but, I guess in my computing life I'm a bit past these adventures and chasing after that last Mhz or FPS.... always cool to see though.

 

[sneekypeet]

Since the OP cannot contain himself to keep the thread on topic.... I'm closing up shop.