[Rumor] ZEN to be "soldered"

[JoeCoke]

45 minutes ago, Pample said:

This is literally the straw that broke the camels back and made me join the Forum.

 

I've been a long time view of LTT and browser of the forum but this kind of annoyed me so much.

 

For what it's worth I am a long time AMD fan (Bring on the torrent of hate mail) and have been since the very VERY early days (Not telling you my age) and I'm truly excited about Zen. Not because of the idea that it might finally bring Intel into check, but rather that it feels for once that us poor AMD fans might actually have something exciting to put into our builds.

 

I really don't care if this rumor is accurate, it's just nice to have something interesting to talk about in regards to AMD, think about it. How long is it since we could really debate something so unknown from AMD in terms of processor? Good and bad, it's exciting.

 

Anyway, to my real point. Rumors are just that and for what it's worth I'm going to go with the APU side, not Zen. Patrick is a troll, he has gained too much attention from all of this (And yes I know how ironic it is that I'm saying this and by proxy giving him more), and removed the focus from something that we should all be excited about. 

 

Phew, rant over.

 

P.S. Hi to all! I'm here to stay now.

 

P.P.S. AMD Rules!!!  

 

P.P.P.S. please direct all your Intel hate to the following address: idontcare@whateverintel.ha.ha

Welcome to the forum.

[Pample]

12 minutes ago, 2Buck said:

Welcome to the forum.

Thanks! Good to be here.

[Energycore]

5 hours ago, MageTank said:

Every single FX CPU I've had, needed to be lapped to make proper contact with my coolers. 

Can you describe the process of lapping?

 

4 hours ago, TidaLWaveZ said:

I'm a bit disappointed guys, I expected this one to get a little crazy. I guess that's what I get for being involved in the most tame, well mannered forum I've ever seen nevertheless been a member of.

It's a forum run by canadians, of course it's well mannered

 

Actually thinking back, when I started getting involved in LTT I thought to myself "man, this forum is nicer than Teamliquid, and that's something"

[Energycore]

1 hour ago, Pample said:

snip

What a triumphant entrance, friend. May your stay here be jolly good.

[MageTank]

Just now, Energycore said:

Can you describe the process of lapping?

 

It's a forum run by canadians, of course it's well mannered

 

Actually thinking back, when I started getting involved in LTT I thought to myself "man, this forum is nicer than Teamliquid, and that's something"

Get a flat surface (glass is ideal), put some sand paper down, and slowly move your CPU across it in one direction repeatedly. That's the basic gist of it. Some people prefer to rotate, others say not to, but as long as you get it nice and smooth (no convex or concave corners) then it's good to go. Some prefer going for the full mirrored finish as well. I even know people that use copper polish after lapping. 

 

Here is what a mirror finish lapped CPU looks like:

 

[Energycore]

Just now, MageTank said:

Get a flat surface (glass is ideal), put some sand paper down, and slowly move your CPU across it in one direction repeatedly. That's the basic gist of it. Some people prefer to rotate, others say not to, but as long as you get it nice and smooth (no convex or concave corners) then it's good to go. Some prefer going for the full mirrored finish as well. I even know people that use copper polish after lapping. 

 

Here is what a mirror finish lapped CPU looks like:

 

Woah that's pretty

 

But in terms of temps, what difference did it make?

[MageTank]

Just now, Energycore said:

Woah that's pretty

 

But in terms of temps, what difference did it make?

Depends entirely on whether or not the IHS itself is fine (flat) or has flaws (convex/concave). My FX 8320 was pretty bad, and lapping shaved off about 4C. It's ideal that you lap the cooler as well, to ensure both surfaces are perfectly flat, but luckily the base of my vapor chamber is perfectly flat (and mirror finished from the factory) so this was not a step I had to go through. 

 

If your IHS is already perfectly flat, you likely won't notice a difference at all. Best way to test, is using a laser pointer and placing a perfectly flat object (i use a razor blade) on top of the IHS. Then, shine a light on the other side of the object, and see if any light comes through the bottom of the object on the opposite side. Move the object several times in different areas of the IHS, and repeat to ensure each corner and the center is flat.

 

Apparently, the guide I used years ago is still up and available: http://www.overclock.net/t/290130/lapping-your-cpu-and-heatsink. Feel free to browse through that thread and look at the results. I didn't lap my current 6700k, mostly because it's already flat according to the laser test listed above. Most of the lapping done these days, are on CPU heatsinks. Mostly air coolers, since most water blocks and AiO coolers use a flat copper slab. Air coolers on the other hand tend to use machined aluminum with copper heatpipes ran through them, which sometimes have uneven finishes that can negatively impact IHS/heatsink contact. Some argue that lapping is no longer viable at all, but I say it's still very situational.

[LucidMew]

37 minutes ago, MageTank said:

Depends entirely on whether or not the IHS itself is fine (flat) or has flaws (convex/concave). My FX 8320 was pretty bad, and lapping shaved off about 4C. It's ideal that you lap the cooler as well, to ensure both surfaces are perfectly flat, but luckily the base of my vapor chamber is perfectly flat (and mirror finished from the factory) so this was not a step I had to go through. 

 

If your IHS is already perfectly flat, you likely won't notice a difference at all. Best way to test, is using a laser pointer and placing a perfectly flat object (i use a razor blade) on top of the IHS. Then, shine a light on the other side of the object, and see if any light comes through the bottom of the object on the opposite side. Move the object several times in different areas of the IHS, and repeat to ensure each corner and the center is flat.

 

Apparently, the guide I used years ago is still up and available: http://www.overclock.net/t/290130/lapping-your-cpu-and-heatsink. Feel free to browse through that thread and look at the results. I didn't lap my current 6700k, mostly because it's already flat according to the laser test listed above. Most of the lapping done these days, are on CPU heatsinks. Mostly air coolers, since most water blocks and AiO coolers use a flat copper slab. Air coolers on the other hand tend to use machined aluminum with copper heatpipes ran through them, which sometimes have uneven finishes that can negatively impact IHS/heatsink contact. Some argue that lapping is no longer viable at all, but I say it's still very situational.

 

 

LTT did a video on CPU heatsink lapping. plus linking a thread on http://www.silentpcreview.com/article1366-page1.html

 

 

[Jovidah]

27 minutes ago, LucidMew said:

 

 

LTT did a video on CPU heatsink lapping. plus linking a thread on http://www.silentpcreview.com/article1366-page1.html

 

 

Up front; I never lapped my CPU nor my heatsink (and I doubt there's a whole lot to gain from it)... and most of my experience using abrasives on metal is from hand-sharpening & polishing my own knives on Japanese waterstones. But I'm going to be nitpicky anway... the way he's doing it sucks.

 

The end-result after his 'lapping' was quite horrible and nowhere near polished. No wonder it made no difference; it really wasn't anywhere close to polished. Loads of scratches, not even anywhere close near a mirror finish. Also a really crappy procedure to do this. 

 

-I wouldn't bother getting every single grit. Yes you need some intermediates to save time but there's no reason you need both a 600 and an 800. You're just adding extra time and steps. 

 

-Doing a specific X amount of laps is...meh... There's really no telling whether X is enough to get out the scratches from the previous grit. Might be more might be less. In his case it seems to have been not enough.

 

-Normally you somehow try to 'test' or measure whether you can move up a grit. There are several sensible ways to do this. The easiest way is to just take a pencil or sharpie, draw a grid on the surface and keep grinding until it's all gone.

An alternative is to not to rotate every 10 laps, but to simply do a certain grit entirely in one direction (only rotate 180 degrees), then do the next grit at a right angle to the previous one. Then you can just resume until you only see scratches in one direction. But that leaves you slightly more open to creating base that might be flat, but creates a bit of a Pisa tower. 

 

To be honest there's probably a good chance it'll still lose slightly more surface on the edges, regardless of how you hold or do it. I also think that no matter how you do it by hand, there's no way in hell you're getting it perfect. Especially for someone who has never done this kind of thing. Maybe with a jig of some sort, but people just have too much wobble in their motor system to really get this perfect.

 

 

Just as a random brainfart; why not put abrasive / polish between the cooler & the CPU and just rub them together? Should make the perfect fit.

[patrickjp93]

On 12/1/2016 at 7:42 AM, Jovidah said:

Yeah I have to admit... I never understood what this patrick vs everyone thing was all about. Until a few days ago I had a 'discussion' with him where he essentially called me an idiot because I wouldn't take his word for it that Kaby Lake would have a 'DMI 3.1 of 8 lanes PCIe 3.0 lanes between the CPU and the chipset'. 

He did actually provide supporting sources for once... except they only refuted his point.

 

He's like the Trump of the internet. Surprisingly bold, stubborn and convinced of himself even when he's talking total crap and has no way to back it up. 

I didn't ask you to take me at my word. I gave you Intel's documentation on the matter you lazy twit.

[patrickjp93]

BTW, for those interested in the truth, It's an Intel CPU in disguise. Intel CPUs DO NOT lay flat against the pads if you put them on a flat surface. They stand a couple millimeters above thanks to the capacitors on the bottom. This is a 65* camera angle with the light source Almost directly above and to the left still creating a sizable shadow. Generating a bit extra to make you think it's standing on pins is child's play.

 

[Jovidah]

20 minutes ago, patrickjp93 said:

I didn't ask you to take me at my word. I gave you Intel's documentation on the matter you lazy twit.

Did you actually read those yourself? Because none of them mentioned either a DMI 3.1 or 8 lanes. 

In fact, Intel never made that claim. Just like I have never seen them claim an IPC increase for Skylake. They're actually quite honest in their marketing...

[ZarosTenjin]

17 minutes ago, patrickjp93 said:

BTW, for those interested in the truth, It's an Intel CPU in disguise. Intel CPUs DO NOT lay flat against the pads if you put them on a flat surface. They stand a couple millimeters above thanks to the capacitors on the bottom. This is a 65* camera angle with the light source Almost directly above and to the left still creating a sizable shadow. Generating a bit extra to make you think it's standing on pins is child's play.

 

still going at it? look at the pcb what does an intel cpu have there (tip its missing something near the corners)  and i dont remember an intel socket that didnt have these nooks (they are somewhere else every time of course) the picture int he OP shows a perfectly square pcb which incidentally is how amd does it look at am4 mbs and the pin socket on those http://www.pcgameshardware.de/AMD-Zen-Codename-261795/News/Sockel-AM4-mit-1331-Pins-bestaetigt-1208073/galerie/2639119/ (its in german but i'm sure you know how to navigate a gallery ). but what is it that riles you so up about this? this is news of the kind of hey thats good or meh whatever but not something bad 

[SpaceGhostC2C]

On 12/1/2016 at 6:51 AM, Jovidah said:

(...)

You focus too much on the details, but he did catch the important part (although only after doing it): the whole point of "direct contact", and ultimately the main reason the 212 got so famous, is the no gaps, lapped flat heatpipes that form the base of the heatsink. Hence, he was lapping a cooler with no curvature and obtaining no gains. He reached that conclusion after the fact, but explained it in the video.

I mean, I don't know if he made a good lapping or not, but the main reason for a negative result was that the cooler needed no lapping to begin with, and that's in the video.

[Dabombinable]

35 minutes ago, patrickjp93 said:

BTW, for those interested in the truth, It's an Intel CPU in disguise. Intel CPUs DO NOT lay flat against the pads if you put them on a flat surface. They stand a couple millimeters above thanks to the capacitors on the bottom. This is a 65* camera angle with the light source Almost directly above and to the left still creating a sizable shadow. Generating a bit extra to make you think it's standing on pins is child's play.

 

Ummm.....you do realise that no CPU to date sits flat? Unless its one that uses BGA?

[Jovidah]

8 minutes ago, SpaceGhostC2C said:

You focus too much on the details, but he did catch the important part (although only after doing it): the whole point of "direct contact", and ultimately the main reason the 212 got so famous, is the no gaps, lapped flat heatpipes that form the base of the heatsink. Hence, he was lapping a cooler with no curvature and obtaining no gains. He reached that conclusion after the fact, but explained it in the video.

I mean, I don't know if he made a good lapping or not, but the main reason for a negative result was that the cooler needed no lapping to begin with, and that's in the video.

Fair point. I guess my main issue with it was 'why make a video that spends several minutes both illustrating and complaining about a bad way to do lapping'. Either do it right or don't do it at all. Especially since his results are so mediocre that they make it impossible to determine whether it actually makes a difference or not.

[patrickjp93]

1 hour ago, ZarosTenjin said:

still going at it? look at the pcb what does an intel cpu have there (tip its missing something near the corners)  and i dont remember an intel socket that didnt have these nooks (they are somewhere else every time of course) the picture int he OP shows a perfectly square pcb which incidentally is how amd does it look at am4 mbs and the pin socket on those http://www.pcgameshardware.de/AMD-Zen-Codename-261795/News/Sockel-AM4-mit-1331-Pins-bestaetigt-1208073/galerie/2639119/ (its in german but i'm sure you know how to navigate a gallery ). but what is it that riles you so up about this? this is news of the kind of hey thats good or meh whatever but not something bad 

Those nooks do not exist on Xeon PCBs. I'm just using this as an example to prove 1 point.

[Ryujin2003]

1 hour ago, patrickjp93 said:

It's not just a possibility. It is fact. And yeah, typo and I forgot E5 vs. E7. Whoopdeedoo. 

 

I don't have to try to appear smart. The only other person on this forum who could match me on both breadth and depth of CPU/GPU/FPGA and programming knowledge was kicked by a mod on an ego trip a couple years ago. The fact you all gang up together to agree on a falsehood just to try to take me down a peg only displays your pettiness and incompetence.

You can believe in whatever truth you want, you are still an ignorant individual. You now have to resort to name calling and group generalization to make yourself feel better?

Since "the only person who could match [you]" was booted from the forum, I guess there isn't any more point in discussing anything intellectual with you. When you and your Napoleon Syndrome decide to get off of the your self-righteous high horse, let the rest of us know.

5 hours ago, patrickjp93 said:

BTW, for those interested in the truth, It's an Intel CPU in disguise. Intel CPUs DO NOT lay flat against the pads if you put them on a flat surface. They stand a couple millimeters above thanks to the capacitors on the bottom. This is a 65* camera angle with the light source Almost directly above and to the left still creating a sizable shadow. Generating a bit extra to make you think it's standing on pins is child's play.

 

PS: Regardless of the light angle, camera angles, wind direction, and weight of a African Swallow carrying a photoshopped AMD/Intel hybrid coconut, this CPU sits much lower than the other one that was originally posted. So Unless Intel gave up on miniaturization, this doesn't explain much.

 

 

[79wjd]

4 minutes ago, patrickjp93 said:

Fine, when an actual delid comes out and looks nothing like that, you can leave LinusTechTips after issuing me a public apology in the off-topic thread. You won't accept evidence right in front of your face SOLELY because it's me supplying it. It's obvious as Monday is long.

I'm not saying it's a Zen delid. But it's definitely not an i7-2830. Oh, and what evidence have you supplied.

 

I'm holding you to the same agreement then. 

@LAwLz @MageTank

[leadeater]

Thread locked.