Etern4l

How do we know the balloon is in fact Chinese, and not say Russian - designed to provoke tensions, or some sort of hoax? Where did it come from? Where is it going? Will it be possible to shoot it down over the ocean? Just some of the questions that spring to mind regarding this ridiculous story...

Fixed it for you sir See, friendship is a strange thing in that it usually requires a good level of understanding, at least one way. Put another way, Linux can be quite friendly, but it's somewhat picky about its friends. I am not sure what you mean by a personal station? Perhaps your requirements are unrealistic? I don't think Linux can clip your nails for instance. More seriously though, while Linux excels on the server, it can these days serve as a great workstation, while beating the hell out of Windows on performance. It can get more difficult to get to work with very recent laptops though, mostly because the user has no say in what components are used. That said, certain laptop lines from Dell and Lenovo (not to mention one or two niche Linux laptops brands) support Linux out of the box.

Actually that kind of makes sense. Alphacool doesn't really specialize in laptop products, and if the paste is very thick, it requires high heatsink pressure, which laptops typically lack. Anyway the paste is holding up so far (in conjunction with a serous cooling upgrade) which is always good.

Same with the MEG ACE, steady progress with each BIOS upgrade, which has been pleasantly surprising, even though it would have been ideal if they had got everything right at launch ;)

Twitter: Number of staff suing goes up daily - lawyer

Rejection of flawed or harmful tech doesn't by itself make anyone a Luddite. Delaying Windows "upgrades" for as long as possible is a time-tested quality of Windows experience maximisation strategy.

Really depends on the config but obviously 1000W+ is easy.

It just takes up 2-3 times the case space, and what if you have more than one GPU - it's a mess.

clkMax=f(T), I think. Yeah, even the 3-way adapters are not great. I replaced mine with an original Corsair cable - so much nicer and more compact.

1200W, oh ye Mighty Bro Fox? This is just for the shunt-modded GPU, right? ;)

So the professional cards can be stacked easily, you don't necessarily have to NVLink them, depending on the application. The reason is they are usually 2 slots or less and don't use as much power. The problem is more with stacking those high-powered geforce cards. Imagine 2x600W 4090 FEs with no space between the cards. Temps?

Yeah, so the question is regarding clearance. I can confirm 2.7 slot cards work well.

The merit is clearly there for some setups, whereas there influencer #1's intro made it sound like it's the best thing since sliced bread overall, which is equally clearly misleading. A fairly minor social media gripe :) rather than anything else - enjoy the build bro.

With all due respect for the bro YT influencer there, as someone who makes system config changes relatively infrequently, I'd rather have the cables tucked in and out of sight deeper inside the case rather than all exposed on the side in a rather unsightly manner and potentially interfering with the side panel. Yes, access is less convenient, but not prohibitively so by any means. BTW I also don't understand the concern about the Type 4 connectors not being spaced out enough in the AX1600i, as a Tom's review claimed IIRC. No issue with that whatsoever. I imagine the side layout could work quite well with an open bench (although would it matter then?), a very wide / cube / non - standard case, or for someone reconnecting stuff on a very regular basis though.

Fair enough, you held out for long enough. The deals on that won't be getting better anytime soon IMHO. Congrats!

Well, ultimately they care about stability, which is where overvolting could help. Anyway, just curious. Trying to eke out more performance from 4 DIMMs and surprisingly getting great initial results with lower than XMP voltage. Early hours of testing though.

Has anyone found that XMP profiles use voltages that are significantly higher than what is actually needed for stable operation?

I highly recommend the Phanteks T30s :more powerful than Noctuas IPPC, but noticeably quieter across the curve (if there is enough clearance to fit them).

I evaluated MX-6 and it's a "not really" for 280W workloads. Testing Alphacool Apex 17W/mK now. It's probably the thickest paste I've ever used, quite difficult to apply, it "peels off" easily, the temps however, were a markedly better than with MX-6 (thanks to @Papusan for the tip). I also have tubes of TFX and XTM70 waiting in the pipeline. My assumption now is just that I will just have to repaste often given the load.

Yes, in the language of gamers and benchers, "ASIC quality" appears to be as unambiguous as it is wrong ;) BTW nein and neun are normally pronounced fairly differently, so context doesn't really matter, unless you are in the Swiss Alps or something lol It's more that the gamers effectively developed a slang term here. If you announced "I've got Ace of Spades" in East London when not playing cards, it would likely be understood that your HIV has sadly progressed to AIDS now, while any non-cockney people present (poker players in particular) would just scratch their heads or start inspecting your sleeves. The example is not quite on point though, because most people know very well what both an ace of spades are AIDS are, whereas the problem here is that gamers don't really know what an the term "ASIC" means. At a risk of an exaggeration, it's like if gamers used the term "quantum superposition" to refer to the "Superposition" benchmark.

A huge stretch. That statement is referring to comparison between different implementations/designs/manufacturing processes. For instance, you could have a cryptominer ASIC designed using different components or implemented using different fabrication processes etc. In the context of "GPU-Z 'ASIC' Quality", we are talking about comparing effectively fairly minute differences in quality between otherwise identical ICs due to the binning process which are all within specification tolerance, and again: GPUs are not ASICs in the first place so that really precludes "ASIC quality" from being applicable. That is just the key confusion here.

Looks like all those links are just in fact referring to "GPU ASIC quality" misnomer, as coined by the GPU-Z author. This could be because the links have been targeted to yourself as the intended audience. Again, the new AI kid on the block beats Google's effort soundly (it does no tagetting beyond the session scope): ASIC quality refers to the level of design and manufacturing processes used to produce an Application Specific Integrated Circuit (ASIC). In general, ASIC quality is assessed based on various factors such as design complexity, performance, reliability, power consumption, and manufacturing processes. Higher ASIC quality typically means better performance, lower power consumption, and greater reliability compared to lower-quality ASICs. The term "ASIC quality" is used to differentiate between different levels of ASIC design and manufacturing capabilities, and to help ensure that ASICs are suitable for their intended applications. Makes sense to me, and is consistent with what I wrote in the previous post (on my own lol).

The term ASIC was coined in the 1980s, I became familiar with it a decade later. Literally all it means guys is that the (AS)IC is custom designed for a specific task, rather than being a standard/general-purpose IC available off the shelf from semiconductor manufacturers. In the past there was much more need for ASICs, because CPUs were so slow and there was no GP GPU computing, but even today, there are applications such that the investment in a development of an ASIC instead of using CPUs, GPUs, standard microcontrollers etc. is justified, usually from the performance perspective. Crypto miners are a somewhat misleading example, although so popular both ChatGPT and bro @Papusan brought it up first, in that it might suggest that ASICs are a recent invention. They are not, and there are lots of modern and historical applications including AI, military, telecoms etc. Once the above is clear, it follows that the use of the term "ASIC quality" referencing modern compute-capable GPUs (CUDA was released in 2006 BTW) is basically nonsensical. 1. The term refers to a design type, rather than a specific type of underlying semiconductor technology and its properties. 2. Modern GPUs are general-purpose off-the-shelf computational ICs rather than ASICs. HTH.

Yeah, planning on that. MX-6 wasn't too bad, but didn't quite cut it. Replaced with Alphacool Apex 17W/mK for now on bro @Papusan 's tip. Clearly much better. I also have Corsair XTM70 and Thermaltake TFX waiting to be tested as well.

What you're essentially saying is that if a sufficient mass of people in their millions repeat something, it becomes the the truth. "Flat Earth! Flat Earth!" :) It doesn't matter what the millions say, it matters what the ones (sometimes very few) who know what they are talking about say. I haven't seen any expert quotes here (Jay is a very nice guy, but clearly not an IC domain expert). Here is one from a dude who's read the all the books and the Internet (OK probably not but close enough): E: Is GPU an ASIC? CGPT: No, a GPU (Graphics Processing Unit) is not an ASIC (Application-Specific Integrated Circuit). A GPU is designed for general-purpose computing and is capable of executing a wide range of tasks, including rendering graphics, video processing, and running simulations. An ASIC, on the other hand, is a type of integrated circuit that is specifically designed for a single task, such as mining cryptocurrency. Not the most in-depth explanation, but the dude got the gist of it. Anyhow, having delivered the message and a few back and forths, I believe I've exhausted the ASIC topic here... It's been quite some time since I studied this stuff, probably getting everything the other way around now ;)