*Official* NBR Desktop Overclocker's Lounge [laptop owners welcome, too]

Your talking to a guy who games at stock. Unfortunately I do not need to over clock to play my games.
As to how much better in this scenario. Do not pay attention to the score, but only to the FPS and percentages for the FPS.
The 2080TI is pretty much maxed out with shunt mods and water chiller while the 3090 is on new vbios and no shunt mods and stock cooler.

This last run was over clocked 10900K@5600 using SS and stock air cooled card.
As to the rest it's hard to say. I would need FE base scores first to have an idea there.

 

Looks like a pretty decent gain honestly. I can see gaming at stock with the 3090 certainly. But running the 2080Ti stock in gaming? I feel like the 2080Ti was watered down quite a bit due to no competition on Nvidias behalf. But, I’m mostly referring to the FE, it runs extremely tame out of the box.

 

Ugh, spoke too soon about silent gaming.
While gaming I heard a very subtle audible "pop" Not super loud or smoke inducing, but definitely a, "Hmmmm, what was that?"
Running a >4 yr old EVGA T2 850w PSU
Within a minute or two, I picked up some coil whine on my GPU which up to that point was running dead silent.
Over the night and today, it has grown louder depending on what I'm doing load wise.
3090 is working absolutely fine, just very audible coil whine.
I used ye ole paper towel cylinder to isolate the coil whine. Dead center of the GPU.
I tried different PCIe cables same result, but moving the cables around to other modular VGA 8 pin plugs on the PSU varied the whine, so I'm thinking it might be the PSU.
Rails look ok, but I had this happen before with a Corsair PSU where rails looked fine, but system was unstable. Swapping in a new PSU and all the little noises and instabilities went right away.
I ordered a 1300w EVGA PSU to test it out since nothing was available locally over 700w.

Oh so that's where you've been! Congrats!

 

Well, to be honest my 2080 Ti holds 144 FPS(hz)@1440P just about maxed out on Warzone and maxed out on Destiny 2.
200 to 500 using the 3090. I'll have to pay closer attention the next time I load that drive up to play.

I have been mainly working for the last few months so haven't had real time to play with over clocking much.

 

That guy is a complete idiot.
First of all you do NOT polish a die or heatsink when applying LM! It simply stops the LM from sticking to the surface and can make it just sit in a ball forever. It also causes it to tend to migrate to itself with time, leaving hot spots when it gets compressed. "Sanding" the heatsink by buffing it with 1500 grit sandpaper to make the surface rough, not smooth works nicely here.

Second, why would you use liquid metal on this GPU? There are so many tiny parts everywhere that make any sort of risks even worse. I would use Kryonaut Extreme or Thermalright TFX.
Far more important than putting LM on the core, is proper use of thermal pads to deal with the hotspots on the chip.

Here are some hotspots on a 3080 Founder's Edition.



Here is a 3090 FE with proper thermal pads (1.5mm) over all of the hotspots and RAM on the front side.

 

2080Ti FE running Port Royal. This Port Royal test crashes at a little more than 380 watt. I did get 10,182 so far. I have never ran port royal before. I’m going to solder shunts on tomorrow, and see if I can get any better. Still on FE air cooler.

I don't think I will have any trouble matching or maybe even out performing a stock RTX3080FE in this testing at all. Looks great so far!

 

I don't really see how shunt mods are gonna prevent crashing when you hit a certain GPU load, shunt mods just underreport the load to the GPU to trick it to go past it's power limitations that are placed in the vBIOS. Shunt modding is likely to make any given overclock you have more unstable if prior to the shunt mod you were bouncing off the power limit, because in this case it would have been limiting clock speed when bouncing off the power limit so effectively at a lower overclock. I mean there's nothing wrong with doing a shunt mod, but preventing crashing I don't think is gonna be one of them.

 

I think you have that wrong. By tricking the card into thinking you are using lower power, you allow the card to draw extra power if it is needed TO BE STABLE. So, if you do not increase the frequency much above the lower frequency when the power limit was the limiting factor, the card is able to draw more power and to be stable at that frequency. You suggesting it is less stable because now it boosts to a higher frequency misses that if you do not allow it to go much higher, it made the frequency before the mod more stable.

It's like saying you took your CPU up a couple hundred MHz after going to manual voltage and it became unstable, all while you are above the frequency you could get stable with the auto-voltage.

Now, there are limits and inherent dangers with shunt modding and everyone must weigh the risk and reward of doing this for their situation (DISCLAIMER). But, I do think your explanation needs more explanation.

 

This is not entirely true.
My MXM GTX 1070 can do +220 core +650 memory at default TDP of 115W.

When the TDP is modded so it can pull 230W, it never hits the power limit and will draw up to its limit without downclocking, usually topping out at 225W peak, with sustained much lower. However with TDP uncapped, I can only do +145 reliably on the core stable. +165 is possible but some things may randomly freeze/crash.

The low 115W power limit keeps the card's clocks low at heavy load, thus allowing a higher offset, which simply makes it hammer the power limit more often, while lighter loads that don't put as much stress on the chip will boost way up.

 

When it goes to or over the 130% limit for too long even a couple seconds it freezes, and causes the benchmark or game to crash. So if the theoretical TDP limit is 170% then it won’t freeze when it is theoretically at 131% TDP or even higher than 131% TDP.

 

The offset DOES NOT MATTER if the raw frequency is higher. For example, let's say normal boost is 1600, and you can add 220 offset. That is 1820 real. Let's say with power limits off, it is a boost of 1750. You then can only offset 145, for a total of 1895, but that is 75MHz higher than before.

You cannot just use offset values. You need to talk in final frequency.

 

I'm fully aware of this. And you're changing the argument to offsets. We're not talking about semantics about offsets here. But you don't set a fixed frequency anyway. There is only base frequency (affected by the pascal temp/voltage curve) and offset to it. But you directly claimed that increasing the power limit lets you use a HIGHER offset. Because "the chip will draw more power."
That's exactly the opposite of what happens on both Pascal and Ampere because the chip runs out of voltage headroom since it is no longer throttling the speeds.
it's almost impossible to talk about frequencies with Pascal because you have to deal with the -13 mhz / 6C problem as well. Makes things hard.

The chip will be stressed more at heavy loads, not at "100%" usage. Usage is directly related to if the chip's core clock is fast enough to allow the chip to do operations to keep it saturated fully. A fully saturated chip needs more voltage to remain stable at higher speeds, *NOT* less voltage. But you run into the 1.062v voltage limit. So to keep you happy let's say this is at 2040 mhz needing 1.062v to be stable (60C). this is assuming unlimited power (no power limit), so let's say the MXM 1070 is drawing 225W absolute *MAX* but averaging 200W, so modded TDP is 230W. It will never hit power limit This is with +160 mhz offset..

2055 mhz crashes because the clock is too high, when pulling 180+ watts (+175 mhz).

Now, on a stock 115 watt TDP 1070, you can now use +220 mhz stable without crashing.
Why? Because the chip will be constantly hitting the power limit and downclocking at heavy load. So it will be running around 1860 mhz instead, at 115W sustained at heavy load. And you can't just lock the voltage in either without doing a lot of work to the curve, because if the target frequency on the curve exceeds the clock the power limit allows, it won't even use that voltage.

So let's say it is running 1860 mhz @ 0.875v and now staying at 115W or below under heavy loads.
So now you add a +220 mhz offset which is 2080 mhz. But it never can reach 2080 mhz. It will only reach 2080 mhz on low load high FPS situations, like loading screens in games or benchmarks, where the card can boost up to push high FPS but isn't put under lots of stress, or in older games that don't utilize the card as hard, allowing it to go to high usage without drawing lots of power.

Since it will never be drawing lots of power at high clocks, thus your higher offset is now stable, since power limit prevents the card from actually working hard enough to become unstable.

 

********. Open GPU-Z or open up any number of logging programs AND MONITOR YOUR FINAL FREQUENCY.

Seriously, offset does NOT MATTER, and NEVER MATTERS, if the FINAL FREQUENCY of the boost is HIGHER. PERIOD. EXCLAMATION MARK.

I said: "I think you have that wrong. By tricking the card into thinking you are using lower power, you allow the card to draw extra power if it is needed TO BE STABLE. So, if you do not increase the frequency much above the lower frequency when the power limit was the limiting factor, the card is able to draw more power and to be stable at that frequency. You suggesting it is less stable because now it boosts to a higher frequency misses that if you do not allow it to go much higher, it made the frequency before the mod more stable."

All of those talk about FREQUENCY. I never said OFFSET! You did. When I say frequency, I mean frequency that the card is performing at, NOT OFFSET.

So you missed the point entirely, and unless you want to talk raw frequency, you are kind of missing the entire point.

But you can ALWAYS FIND THE AVERAGE FREQUENCY and you can always compare that to before your mod.

Edit: Let me say this another way. At the 180MHz frequency delta between the throttle and the 2040MHz, you would have to raise the card by 83C or so to cause that much throttle, which means you are really at like the thermal throttle point of the card, causing a hard throttle. If you are allowing yourself to be heat constrained, you should NOT be doing the mod in the first place, especially with the example you gave.

Instead, you should be on water or colder.



In other news:

 

Do you even know who I am dude? You're acting like I'm a beginner.
I *shunt modded* my 3090 FE and TDP modded my GTX 1070 and hacked the power limits on my MSI laptop.
I know full well what I'm talking about and my observations are airtight.
I have absolutely no idea why you're acting like I'm a beginner.

I'm done with this conversation.

 

I treated you like a person that ignored my words to say I argued something that I did not. Your first response falls into a clear issue that others fall into on the internet, which you did not fully clarify in your second post. Do I think you know the difference? Probably. Did you clarify it in your longer statement? No. Did I find you not giving me the initial benefit of the doubt to tease out what I was saying disrespectful, as you see my response to you? YES.

So, let's start over at the top to lay out what each other were saying, the problems each of us have with what we said, and figure out where the tension is instead of acting like children.

1) I was discussing SOLELY the effect that removing the power limit can have. That included the theoretical part of being able to keep other factors equal. You are correct that removing one throttle event can lead to other throttle events. I do not think I ever said it would not. So, that is one problem.

2) When you responded, you only discussed purely in terms of offset. This did not include the baseline reference frequencies BEFORE and after. You then talk of the offsets again. Offsets are not the end goal. Higher frequency is. And others that know less than you literally have argued that they get less overclock after, when having higher end frequency. So, let's spell out more what we are saying, because if I re-read your statement correctly, you did not fully contradict me.

You said:
"2055 mhz crashes because the clock is too high, when pulling 180+ watts (+175 mhz). [discussing the variant with no power limit]

Now, on a stock 115 watt TDP 1070, you can now use +220 mhz stable without crashing.
Why? Because the chip will be constantly hitting the power limit and downclocking at heavy load. So it will be running around 1860 mhz instead, at 115W sustained at heavy load."

You also acknowledged "you can't just lock the voltage in either without doing a lot of work to the curve, because if the target frequency on the curve exceeds the clock the power limit allows, it won't even use that voltage." Now, my statement was simple, that "(b)y tricking the card into thinking you are using lower power, you allow the card to draw extra power if it is needed TO BE STABLE." So you are correct that this, outside of a vacuum and outside of theory has the effect to now allow the card to boost higher, potentially into unstable territory. You are also correct that you would have to do a lot of work on the voltage curve to try to clamp things down. The problem came when you assumed I said one thing, when I was making a more general point, then you point to other issues, which are real issues, that arise after taking away the power limit.

Inherent in your example, there are higher frequencies AFTER the mod than before. You are correct, and I do not say you are wrong, that pushing beyond a certain frequency causes instability. What you clearly say is that you could use a 220 offset to get to 1860MHz before the mod, then you cannot use 220 after, as you are limited at 2040 or else you hit instability. But that is still nearly 180MHz above the stock behavior. So, looking at playing with the power curve and target frequency, you can make less than that stable, giving you stable at the old peak up to like 2040 without too much work.

A necessary part of all this is that you did the mod to remove the power limit to be able to achieve a higher frequency under at least some circumstances, otherwise people would not do it. Can we agree on that?

Some noobs, not you, think that because you have a lower offset after the mod, you have a lower overclock. But, if the frequency is now boosting higher without applying any offset after the mod, you do not need as large of an offset. That is what I meant in this example not based on real world values: "The offset DOES NOT MATTER if the raw frequency is higher. For example, let's say normal boost is 1600, and you can add 220 offset. That is 1820 real. Let's say with power limits off, it is a boost of 1750. You then can only offset 145, for a total of 1895, but that is 75MHz higher than before."

What you gave in response, as you said "[you] know full well what [you're] talking about and [your] observations are airtight," is an example from real world experience regarding hitting other limits after the fact, while also trying to give a nuanced discussion regarding boost frequency. It is well worded and is based on personal observation outside of theory. But, it also is missing what I just mentioned on whether the final achievable frequency is higher, or rather if you can stabilize more easily the prior highest frequency after performing the mod to remove the power limit. That doesn't mean your observation is wrong, it means it isn't addressing my point. Although you are correct that the offset value will be lower for the reasons you state, and you are correct that you may hit other throttle events or limits, such as voltage.

3. Did I need to be such a prick over the mischaracterization regarding offsets versus frequency achieved? No. And for that I can apologize for and I am sorry for. I also used an asinine example regarding the heat/frequency curve where more heat reduces the frequency, ignoring the boost behavior's full accoutrements of factors in behavior. That, too, I should apologize for and can say sorry for.

So, I hope this explains why we talked past each other, specifically point 2. No one likes to feel unheard, be challenged, be called wrong, or be misunderstood. Although related, we were making slightly different points and I feel we were talking past each other, rather than talking to each other. I apologize for the behavior I feel I should be sorry for. If I missed something, let me know. I also hope I explained how you are correct in what you said, but just slightly missed the point I was making, just as I made the mistake of not acknowledging your detailed discussion on boost behavior and belittling your proffered observations. That was wrong of me when I very well could have tried writing something like this in response.

 

Got a x170sm with a 2080s hoping to see if anybody has done any mods that maybe interesting. If so could any point me in the right direction? I usually see Gods like slv7 and @Johnksss doing these thing but haven't seen him or slv7 in the x170sm lounge so came here.

 

My guess is your overclock is not stable at higher current draws, shunt modding is not gonna change that - that's my intuition on it.

 

Yep, I have the same experience, if I artificially limit the TDP on my GTX 1070 then I can run a higher offset overclock. (Yeah, shunt mods won't increase overclock stability of any given offset overclock, if anything the opposite).

 

Does it boost higher without any offset than the prior boost without the offset? Is the final frequency higher with the shunt mod than without, regardless of your offset value?

edit: Here is an image to show what I am talking about:


Notice how after applying the first and second shunts, the frequency was higher. If that is higher, then your offset plus the higher frequency is likely higher than not performing the shunt mod and using a higher offset.

 

I don't think I'm totally understanding what you're asking me, but my GTX 1070 doesn't need a shunt mod as it has a silly high TDP limit anyway of over 260W if I remember rightly, so it's never bouncing off the current limit....I was talking about when I lower the TDP value below 100% - that allows me to run a higher offset overclock, of course though it doesn't result in higher clockspeed. My first point to tps was that I'm fairly certain his crashing when he reaches a certain current amount is likely due to higher current draws being more testing of an overclock's stability - therefore him applying a shunt mod can only really make a particular offset overclock more unstable, not more stable, or if he's not bouncing off the power limit then it will do nothing to make it any better or worse in terms of stability. I just think he's barking up the wrong tree expecting a shunt mod to increase his overclock stability, I can't really see it any other way based on my experiences & understanding.

 

Our argument started when I mentioned, if power limit is your limiting factor, your overclock at a specific frequency would become more stable, excluding other potential limits on the card, such as thermals, voltage, etc. He took that to mean I was talking of overclock offsets. Off- down the rabbit hole we went.

Your post talking of observed behavior, then my response, was to attempt to get an outside to, in part, mediate where we had the issue of disagreement only on a specific point of contention.

That is the reason I asked on the final frequency. Also why i posted Der8auer's table of results showing shunt mods raising the boost frequency before applying an offset.

I do apologize for attempting to drag you into this.

 

Brother Fox, I don't know how to thank you for this. Puts a lot of confidence in Intel platform build for Win7 lol. Have been too busy with all the usual stuff, was browsing today casually and thought about the new X570 Crosshair VIII Dark Hero from ASUS, and I came across a video on X570 Win7 installation. That new Dark Hero board has passive X570 chipset cooling like X570 Xtreme and a Realtek LAN probably has Win7 drivers which work too, as for WiFi AX200 same boat. It apparently has CSM option too, TPU got the BIOS Screenshots.. in their review.



Even with LTSC it's Win10 is hell due to WDDM mandate and the new LTSC is coming soon as well. God knows what they will enforce on the WDDM requirements for the new games (not much interesting to me nowadays esp with some sociopolitical crap in them) however some gems are great - Mafia Definitive Edition, DOOM Eternal, DMC V, Halo MCC, RE2R all work on Win7 btw. Plus will dual boot anyways still would be great to have a good old solid installation for Win7. Ngreedia Ampere supports Win7, but AMD RDNA2 abandoned it.

Also with the new Ryzen 5000 AMD seems to be a great option, however RKL probably will match it or beat in gaming, I have serious doubts on the SMT area Ryzen is just chomping down in that dept. and RKL also going to have the X570 Chipset I/O parity due to the DMI link having more lanes this time with Z590.

On DDR4 and both these Z590/X570 will probably be the last of the machines which will be able to run Windows 7. After that new Alderlake whatever biglittle drama bs Intel is going for and new DDR5 from AM5 and next Nvidia chip, I think Win7 is going to fade away soon.

Wonder if anyone here tried to install Win7+X570 here..

 

Sadly Doom Eternal no longer works on Windows 7 for me since the game's Update 1 in May. Performance absolutely tanked with wild fluctuations in FPS and frametime and it locks my display at 60Hz with massive screen tearing. The game runs fine on Windows 10. Also the D3D12on7 games I tried either didn't work or had weird game-breaking performance issues, like being unable to select exclusive fullscreen mode or frame doubling (actual unique frames are half of what the FPS counter shows). Running the same games on Windows 10 in native DX12 didn't have these issues. Sadly I think Windows 7 is obsolete as a gaming OS for modern titles. I doubt developers even bother to test their games on it for issues anymore.

 

Yeah, they're really such manipulative bastards. I feel sorry for the people that rely on employment with such a vile outfit as the Redmond Reprobates. They are truly worthy of all of the hatred they receive as a company.

 

And it turns out it is remote desktop protocol that is one of the issues I have causing win 10 to keep saying you have an upgrade even with the most recent upgrade already incorporated into the image.

https://blog.netop.com/what-to-know-about-rdp-security

https://cyware.com/news/understandi...-desktop-protocol-over-the-internet-44351323/

https://www.cloudflare.com/learning/access-management/rdp-security-risks/

https://www.passportalmsp.com/blog/security-risks-of-remote-desktop-access

https://www.darkreading.com/endpoin...top-access-are-far-from-remote/a/d-id/1331820

https://blog.malwarebytes.com/exploits-and-vulnerabilities/2020/10/brute-force-attacks-increasing/

https://en.m.wikipedia.org/wiki/BlueKeep

https://blog.rapid7.com/2020/10/13/patch-tuesday-october-2020/

https://nvd.nist.gov/vuln/detail/CVE-2020-16896

That's right, last month it was so bad the government highlighted Microsoft and their patch of RDP. Everyone should gut it from their builds, along with remote registry and remote workstation and remote assistance. If you don't have to have them, then you do not need the risk.

But, if you try removing RDP from server 2019, it borks the build entirely, making it unusable. You can easily remove it in NTLite. You can, in a way more complicated way, so so with powershell and enough knowledge with DISM and other Microsoft deployment tools.

You also need to make sure the setting is turned off to allow sync on network to not explicitly paired mobile devices. Why would Microsoft even allow for such a thing?

But I digress. Security first. So if you must use windows 10, GUT IT! LTSC, by itself, it's not enough.

 

EK Launches QuantumX Delta TEC Water Block ekwb.com | Today

It is expected to ship out in early December

EK-QuantumX Delta TEC - Thermoelectric cooler for Intel ekwb.com | Today



Pre-Order €294.03

 

I just saw that. $359.99 USD. Just need a 200W overage for your PSU to be able to feed it, or abouts, and an 8 pin.

Wish they had something like this for AMD, both mainstream and Threadripper, which also would work with the 3175X from Intel.

Edit: Here is a 360 all in one if you prefer that-
https://landing.coolermaster.com/pages/ml360-sub-zero-cryo-clocking/

Edit 2: Here are LTT and Der8auer's videos:

 

Linus gets it.... (re: 3090 coil whine)



I returned a GTX 1080 years ago for the exact same thing. I can't do coil whine. I know it is a normal thing and some cards have it, but my ears are hyper sensitive and it drives me batty. I usually don't wear headphones either.

 

Can't find where to purchase the 360 AIO unit, I am going to grab one for more overclocking fun. I have a pretty golden 10900K myself but need thermal headroom. My 1600w T2 PSU was made for this thing lol.

https://videocardz.com/newz/nvidia-geforce-rtx-3080-ti-to-launch-in-january-2021

Also looks like 3080 Ti coming Jan 21' for $999 with 20gb VRAM. Nice! Will replace my 3080 with it.

 

I've got the same PSU and a G2 1600W, so I hear you. I don't know where to pick one up yet. EK has that waterblock on pre-order.

As to the 3080 Ti, I'm going 6800 XT. $650 fits my budget better and gives the money for a waterblock.

 

I had a GTX 275 that did that, quickest sale I ever made.

 

My new 1300w PSU arrives today. If it is the card and not a dying/weak PSU, I'll be selling it or returning it. There's no way I'm dealing with coil whine for 2-3yrs.

 

Does the EKWB Cryo fit X299?

 

No, this is LGA 1200 only unfortunately. Also sounds like Intel is bringing new overclocking features to 10th and 11th gen via a new Turbo boost or allowing manual adjustment of TVB which can actually be useful as those cores are your best on 10th gen and are identified in the BIOS.

 

It is technically only Intel 1200 pin compatible.
https://www.ekwb.com/shop/ek-quantumx-delta-tec

Linus said you might technically be able to run it in unregulated mode on anything, which then I wonder if the footprint is enough for the x299, but doesn't seem like it initially.

There are some bios settings that are needed, etc.

 

Me neither, I use headphones and it still drives me insane.

 

isn’t coil whine a normal function of a GPU when running unrestricted FPS at low resolution?

 

Yes. Turning on Vsync normally makes the noise go away, or significantly reduces the volume.

 

Not on that card, it only changed the pitch, not the dB

 

You would need to mod it to use on X299, Ryzen or bare die 10th Gen as the bottom of it is designed to fit over the ILM (unfortunately). But, this seems like it could do the same job as a water chiller for half the cost. I like it and I think I want one. It would be perfect if it worked with a bare die. I wonder how cold I could make things using it with the water chiller since the temperatures are very similar? Or, maybe I can sell my water chiller and free up a little space in my office.

 

I run @ 4k and capped at 60fps (display limit), G-sync enabled.

Card runs great, but the coil whine is going to force my hand. I know most don't care, but it is a deal breaker for me.

To add insult to injury.....

I received my EVGA 1300w PSU from Amazon today and it is clearly used and clearly DOA.

 

Unfortunately, if the ML360 TEC performance is anything to go by, I think you would be very limited compared to the liquid chiller.



As you can see, the AIO starts being about the same with the regular AIO around 230W, and any HEDT Intel chip clocked high enough can blow by that. Hell, my 1950X would say hello, goodbye under load.

The TEC used is only large enough and tuned for a lower wattage power solution. So modified it could work great with Ryzen 5000 systems, funny enough. But, for our beefier toys, we'd have to swap out the TEC peltier for a larger one, then retune the controller for use with the larger peltier, while throwing another 6 pin or so on it to feed it, etc.

But if EK and CoolerMaster are making this for Intel's comet lake and upcoming rocket lake (8-core), if enough people grab them, then it may spread.

This really is one of the best implementations of TEC I've ever seen. I forgot which boutique system builder a couple year's ago used a similar solution. Either way, they stopped after a generation or two. But they had a similar solution of baseplate, TEC, coldplate for the liquid cooler, and top/hoses going to a custom loop. I just cannot remember their name, but I liked it.

Didn't have the control on the TEC that this does, though, as that TEC was on all the time and had a different control solution, IIRC.

In any case, wait to see if they get a beefier one before making the switch. We are still limited by the TEC they used, and using the water chiller on the hot side, although it helps for efficiency, you become limited before you will get the most out of it, and you may even overpower it and run hotter than with the chiller.

But, still REALLY cool of Intel to help design this. Their engineers are good, always said that. They just hit some issues on manufacturing process and that backed everything up, then instead of backporting, they pushed forward, then by the time they changed course, the entire industry caught up on process or passed them. On microarchitecture, they technically have designs where Zen 4 and Golden Cove should have similar IPC. Just Alder will be on 10nm and Zen 4 on the new 5nm. So process will play a role. But rumors do have Alder running faster than tiger, which topped at 4.8GHz single core stock, so that IS good news. Just a question if AMD will keep having increases in speed like they have been.

Seems like we really are getting into that good stuff coming up!

 

I think this new cooler is best for people who mostly use their systems for gaming rather than those who have lots of productivity workloads.

Paired with an i5 10600k or an i7 10700k, you could overclock the crap out of them for the best possible gaming performance they can offer. That's the ideal use case for a TEC.

It'd be nice if TECs weren't super inefficient, because they require craploads of power to work. I'm hopeful the technology will further improve over time.

 

It's only been almost 200 years since Jean Charles Athanase Peltier discovered electricity between dissimilar metals would cause the production of heat/cooling. Let's give it some time... Hah but in all seriousness it would be fun to play around with one of these...

 

I think a water chiller would do better for me. My CPU just puts out to much wattage. It is unfortunate that it only fits a 10900K though..

 

I think you can pull off that rubber shroud. It looks like it only slips onto the bottom.

 

https://www.3dmark.com/3dm11/14113421 | https://hwbot.org/submission/4599681_mr._fox_3dmark11___performance_geforce_rtx_2080_ti_45418_marks

 

Nice! I see you changed out your motherboard, how do you like the EVGA Dark compared to the Apex (I think you had an Apex).

Also nice to see you back here, I was wondering where you went.

 

Does the block fit the thread pattern on X299 though?