Terribly Misleading LTT Video

[flibberdipper]

I dare you to run a business like what Linus does and then try to make all of your videos of the same quality when something less-than-ideal happens, or just try to make good videos 365/24/7 for that matter. It won't happen.

[handymanshandle]

I would have to disagree with you (OP) because for what delidding is worth there's not much that makes all the effort worth it.

 

The stock cooler being used is fine. Now I would see a big problem if he was using a variety of coolers with out any consistency.

 

 

[Coaxialgamer]

Just now, wcreek said:

I would have to disagree with you (OP) because for what delidding is worth there's not much that makes all the effort worth it.

 

The stock cooler being used is fine. Now I would see a big problem if he was using a variety of coolers with out any consistency.

 

 

I disagree . In overclocked setups , it isn't unusual to see a 10-20C drop in temps . But if it is worth it is up to the person doing the delidding

[-BirdiE-]

22 minutes ago, tmcclelland455 said:

I dare you to run a business like what Linus does and then try to make all of your videos of the same quality when something less-than-ideal happens, or just try to make good videos 365/24/7 for that matter. It won't happen.

The reasons this argument is invalid have already been stated in previous comments.

 

2 minutes ago, wcreek said:

I would have to disagree with you (OP) because for what delidding is worth there's not much that makes all the effort worth it.

 

The stock cooler being used is fine. Now I would see a big problem if he was using a variety of coolers with out any consistency.

 

 

The stock cooler being used is not fine. If you think it's fine then you don't understand heat transfer.

[MageTank]

14 minutes ago, tmcclelland455 said:

I dare you to run a business like what Linus does and then try to make all of your videos of the same quality when something less-than-ideal happens, or just try to make good videos 365/24/7 for that matter. It won't happen.

True, but I wouldn't make the same rookie mistakes that he made in this specific video. He treats delidding as if it's something done just to lower overall temps by a few degrees, when in all actuality it's done for one very important reason. Sometimes, there are air bubbles on the die itself, and certain cores can run 10C hotter than others. That was my #1 complaint when watching this video. He didn't acknowledge that fact. Even if you watch the video yourself, you can see before the delid that his CPU cores temps are:

Core1: 70

Core2: 70

Core3: 60

Core4: 63

 

That is a 10C difference in core temps. After the delid, his core temps are:

Core1: 59

Core2: 65 (with spikes to 72C, possibly due to him not spreading, and just hoping to have even mounting pressure on the IHS, thus an air bubble formed)

Core3: 61

Core4: 65

 

However, the results speak for themselves in the video. The difference in temps between the cores themselves went from 7-10C on average, to 4-6C (with a spike to 13C difference, again, possibly due to the reason i stated earlier). That is why we delid. Letting one core run drastically hotter than the others can result in the rest of your cores being stable at a certain OC, but others to fail. Having the core temps as close together as physically possible is just better, period.

 

Also, the LTT video quality did not degrade after moving. They have had... questionable videos in the past. That "memory speed for gaming" video still haunts me. I mean, man, that video was all sorts of fail. GTX 660 Ti running settings WAY outside of its league, with quad channel ram that is actually slower than my dual channel kit? Still cringe over that one. 

[iamdarkyoshi]

Speaking of delidded CPUs, is the core2duo e8400 soldered or not? Because I get pretty large temp differences, even idle. Mind you, idle is like 50 degrees and load is 80

[.:MARK:.]

2 hours ago, bob345 said:

A company like the one Linus is running is a whole lot more than just making videos. Thats what people dont seem to understand. There is SO much more to it than that. Anyone who has ever had a job or started their own company will know that. What you see on the surface is usually not even 10% of what is actually going on behind the scenes. Also, companies like evga will have the resources to do a move without any repercussions because the company is so large. This is not true for a small company. A move can either make or break a small businesses. Go get some real world experience before you go out and say somthing is simple and "straight forward" because its not.

Do you understand it? Because what you're saying is just pure common sense, it doesn't sound like you know much more than anyone else in this thread. I'd have to agree that the recent content seems less researched and planned. And just because the company may be operating in "less-than-ideal" conditions, does not mean that we can't/shouldn't criticize the content produced.

 

14 minutes ago, tmcclelland455 said:

I dare you to run a business like what Linus does and then try to make all of your videos of the same quality when something less-than-ideal happens, or just try to make good videos 365/24/7 for that matter. It won't happen.

Poor response honestly, I can't understand how you can say that. Linus isn't better than us, he is an ordinary human being, why do you put him on such a high pedestal? Is it impossible for him to have made a mistake and produced content that people aren't happy with? Because if he did, I'm sure he would want to rectify that as he is trying to produce content that people will watch and hopefully enjoy.

[QueenDemetria]

1 hour ago, MagnesiumPC said:

Sorry, haven't seen the video. (I don't support Vessel).

I'm guessing that he has at least some sense and controlled his variables.

If he doesn't, this is unacceptable.

Please let me know.

It wasn't exactly scientific, so first of all ambient temp wasn't measured at any point during the test. My next little gripe(which probably doesn't make a difference, but still is a flaw in the test) is letting the thermalpaste sit for a full day before testing the baseline, while the delided CPU was tested right after installation. This one kills the results though: they did not monitor fan RPM or use a static RPM, so this is yet another and very crucial variable, which could have given the results that we saw.

[Coaxialgamer]

2 minutes ago, iamdarkyoshi said:

Speaking of delidded CPUs, is the core2duo e8400 soldered or not? Because I get pretty large temp differences, even idle. Mind you, idle is like 50 degrees and load is 80

yeah , all pre-ivybridge cpus are soldered

[iamdarkyoshi]

Just now, Coaxialgamer said:

yeah , all pre-ivybridge cpus are soldered

Except SOME pentium 4 CPUs I can delid just by removing dell's heatsinks because the damn things are stuck on so hard

[handymanshandle]

4 minutes ago, -BirdiE- said:

The reasons this argument is invalid have already been stated in previous comments.

 

The stock cooler being used is not fine. If you think it's fine then you don't understand heat transfer.

Yes a better cooler would probably be representative of the kind of people who'd consider doing something like this.

 

If you're trying to oc something that can't run stable and at a reasonable temperature your problem might not be the IHS it could be simply too much voltage or inadequate cooling.

 

But yes, I get at least the basics of heat transfer.

[Coaxialgamer]

Just now, iamdarkyoshi said:

Except SOME pentium 4 CPUs I can delid just by removing dell's heatsinks because the damn things are stuck on so hard

really ? How does that not destroy the die ?

[iamdarkyoshi]

Just now, Coaxialgamer said:

really ? How does that not destroy the die ?

These ones did not have a soldered IHS. They were back when intel believed in pins on the CPU

[Coaxialgamer]

Just now, iamdarkyoshi said:

These ones did not have a soldered IHS. They were back when intel believed in pins on the CPU

what pentium 4 was it ? prescott ? Northwood ?

[iamdarkyoshi]

Just now, Coaxialgamer said:

what pentium 4 was it ? prescott ? Northwood ?

 I would have to look. I used to have a whole shitload of them

[WkdPaul]

Just now, iamdarkyoshi said:

 I would have to look. I used to have a whole shitload of them

same here, I remember I delided my P4 1.8 and ran it @3.2 for 2 years before it died! lol

[-BirdiE-]

2 minutes ago, wcreek said:

Yes a better cooler would probably be representative of the kind of people who'd consider doing something like this.

 

If you're trying to oc something that can't run stable and at a reasonable temperature your problem might not be the IHS it could be simply too much voltage or inadequate cooling.

 

But yes, I get at least the basics of heat transfer.

The issue is that the stock cooler was potentially a bottleneck in the thermal transfer, so even if he delided and it made a HUGE difference in the die being able to transfer heat to the IHS, it would still be limited by the cooler's ability to dissipate the heat and would show as not a big difference in temperature. By using a much better cooling solution you could remove any chance of the cooling solution being the bottleneck and making your results invalid.

[.:MARK:.]

2 hours ago, MagnesiumPC said:

If you don't like the content, don't watch it.

Here's a scenario:

 

Person X likes a YouTube presenter and his channel, and has subscribed.

Person X notices that recent videos aren't great, they may be rushed or poorly researched.

Person X criticizes the recent videos on the community forum, hoping they get better, so he and the presenter can both benefit.

Person X meets an aggressive member that implies that he should go elsewhere, generally being disruptive.

 

Person X does nothing wrong here.

Figure out who is.

[BrightCandle]

I took issue with the CPU thermal paste tests as well because they really did a poor job. There are plenty of good guides to thermal paste showing that different thermal pastes need different application methods due to viscosity differences and that different heatsinks need multiple mounts as the quality of the contact can differ. But what we got was a very basic  video that didn't even reach the state of the art and ended up coming to a conclusion that many know is wrong for many many years.

 

The reality is LTT just doesn't do the research and you very quickly out grow a lot of their content, its basic at best and often misleading. Its also a large amount advertising, not just directly but also "native" advertising where products are placed due to sponsorship deals and you just see the products in all sorts of videos and never realised they were paid to use it.

[handymanshandle]

2 hours ago, -BirdiE- said:

The issue is that the stock cooler was potentially a bottleneck in the thermal transfer, so even if he delided and it made a HUGE difference in the die being able to transfer heat to the IHS, it would still be limited by the cooler's ability to dissipate the heat and would show as not a big difference in temperature. By using a much better cooling solution you could remove any chance of the cooling solution being the bottleneck and making your results invalid.

I guess I could see that being a problem. Though I'm not completely sure how indicative a cooler would really show significantly different results though maybe the cooler is the bottleneck and delidding might actually have a noticeable benefit. 

[LAwLz]

5 hours ago, Stuff_ said:

Yup, Linus is assuming the heat transfer is a linear curve, and you'll see the same results (3-4c change) no matter what temp you had prior to delidding.

 

I don't believe that is the case, and I think it is entirely dependent on the cooler.

You would be right in believing that.

 

It's the same with testing coolers with stock clocks. What might be a 10 degree difference at stock temps might be 15-20 degrees overclocked.

In fact, I am fairly sure they made a video before showing that as you increase the clock speed, the gap between coolers becomes larger and larger. It does not scale linearly at all.

[Bhav]

I just had a quick research and fortunately the 5820k I'm getting soon is soldered.

 

Another plus point for X99.

[flibberdipper]

On 2/26/2016 at 5:14 PM, .:MARK:. said:

Poor response honestly, I can't understand how you can say that. Linus isn't better than us, he is an ordinary human being, why do you put him on such a high pedestal? Is it impossible for him to have made a mistake and produced content that people aren't happy with? Because if he did, I'm sure he would want to rectify that as he is trying to produce content that people will watch and hopefully enjoy.

How the hell am I putting him on a pedestal? Last time I checked, comparing some person on a forum to a guy running a business was not putting said business owner on a pedestal.

[Prysin]

On 27/02/2016 at 2:13 AM, LAwLz said:

You would be right in believing that.

 

It's the same with testing coolers with stock clocks. What might be a 10 degree difference at stock temps might be 15-20 degrees overclocked.

In fact, I am fairly sure they made a video before showing that as you increase the clock speed, the gap between coolers becomes larger and larger. It does not scale linearly at all.

@LinusTech

I think we can add in todays Vessel video about "does watercooling heat up your room".

Which has ONE major issue.

 

you used an open test bench, where ALL the heat is dumped into the room. Whilst if you used a case, like say the Mastercase which we all know you got, it would leave SOME of the energy trapped inside the case.

 

It is common knowledge that IF you got a case, air cooling with generally dump less heat into a room IN A SET amount of time. Ultimately, the same amount of heat MUST leave the case in order to cool a component, but that does not mean the impact on the room will be as pronounced as with watercooling where the heat is actively dumped outside the case and not inside it.

Energy must transfer, but the rate at which it transfers under real life conditions (aka, having an actual case), will be notably different depending on the method of transfer.

[-BirdiE-]

1 hour ago, Prysin said:

@LinusTech

I think we can add in todays Vessel video about "does watercooling heat up your room".

Which has ONE major issue.

 

you used an open test bench, where ALL the heat is dumped into the room. Whilst if you used a case, like say the Mastercase which we all know you got, it would leave SOME of the energy trapped inside the case.

 

It is common knowledge that IF you got a case, air cooling with generally dump less heat into a room IN A SET amount of time. Ultimately, the same amount of heat MUST leave the case in order to cool a component, but that does not mean the impact on the room will be as pronounced as with watercooling where the heat is actively dumped outside the case and not inside it.

Energy must transfer, but the rate at which it transfers under real life conditions (aka, having an actual case), will be notably different depending on the method of transfer.

Going to have to disagree with you on this one.

 

Pretty much any case these days has good enough airflow that the amount of heat "held" by the case would be negligible. The only reason that the air coming out of the case is much hotter is because it is right beside the source and the heat hasn't had enough time/space to mix with the cooler air of the room.

 

Basically the principle Linus was getting at was this:

Let's say you have a space heater with a heating element and fans blowing over it... Assuming the heating element is a constant, if you blow the fans over it at 50% speed the air coming out of the heater will be much hotter than if you have the fans blowing at 100% speed, but there won't actually be any additional heat in the room. This is because, when the fans are blowing at 100% they have half the time to pull heat away from the element, but it heats double the volume. What the faster fan speed DOES do is move the heat is does pick up away from the element faster, causing the physical heating element to be cooler while outputting the same amount of heat.

 

While you're technically correct in saying that, in a closed case, it would take longer to remove the heat from the system by having the heat contained for a bit longer, the effect that would have on a 3.25 hour test is incredibly small to none. Let's say it takes two seconds for the air to enter the case, go over the hot components, and exit the case. That means at any given time the case is holding 2 seconds worth of heat output in a test that is 11,700 seconds long. Doing some quick math... If the case is holding an extra 2 seconds of heat, and the temperature raised 6 degrees over 11,700... Then if he had a case for the test the room would have been 0.001 degrees cooler.