Jump to content
Search In
  • More options...
Find results that contain...
Find results in...


  • Content Count

  • Joined

  • Last visited


This user doesn't have any awards

About pr0xZen

  • Title
  1. I've seen people talking and asking a good amount of questions about these aspects lately - I'm guessing both caused by the improvements to creative workload performance-value with Ryzen 3000-series and the sheer brute force with now decent clock speeds, of Threadripper 3000-series. And, how does it fare without hw accellerated [en/de/trans]coding. So I'll try to fill in a gap, as this thread is already well informative and not too bloated. By my personal experience - yes. At least with Adobe and suitable render loads. While its most recent iterations makes significantly better use of both multithread and GPU accelleration than historically, GPU usage is still very often in the 20s to low 30%s. So a superbeast GPU instead of a solid but more humanely priced GPU, might not actually make as big a difference as quicksync can/will, as the GPU isn't taxed anywhere near its potential. Where quicksync hw accelleration is available and supported, it will still make a very decent contribution. It will not make ie. a 9900k outperform a 3960X (or even a 3950x) at Adobe rendering, but as to your question; there will still be quite notable gains with quicksync, everything else being equal. For H264 that is. With other workloads and setups, YMMV of course. In regards to Intel and any offering desktop SKUs physically without an iGPU; Economy of scale, and profit margins. Sales pricing of CPUs don't really scale with physical silicon size, not at these die sizes vs volumes. The vast, vast bulk of sales pricing is recuperating RnD and overall operations expensed, and profit margin. Die physical size is technically and financially relevant, but at the consumer end of the transaction, price is what people are willing to pay - usually near or at the pain threshold. Price is somewhat arbitrary, and Intel would never cut pricing to any *substantial* degree, for "the same" CPU sans iGPU. The math on this is likely to change ovee the next few years, as cost per die and wafer is posed to go through the roof with the coming (and in-work) smaller nodes.
  2. Thank you for the kind feedback. I've always been a big fan of creative, DIY mods, and never shied away from bringing out the powertools and breadboards/protoboards. So the moment I saw this little gem, modding thought kicked into overdrive hard I now have like 6 of them lying in wait for future mods and ideas. This has been my first custom WC loop/setup though, a perfect project for the tinkering modder. I was aiming for a pick-and-choose AliExpress loop myself, but got reeeally lucky on a couple of 2nd hand market deals, so the core components of my loop are EKWB, but I ended up around the same total cost estimate. Only significant change of design plans was going from 2x360 AliExpress rads to 240+480 EKWB rads. Took a bit of cutting to shed the optical bay cage and arrange secure mounting, but it ultimately made space for 480mm EK rad at front intake. Ample room, at least 2-3mm to spare in my Obsidian 750D! Well all core parts EKWB except for the very recent 1080ti block, which i got still sealed with my 2nd hand purchase of the GPU. Up until a few weeks ago I was running 2x980ti with EK fullcover blocks. I do have a bunch of fittings and smaller parts from Ali though; flow indicator, 45° and 90° swivel fittings, extenders, adapters, plugs, plugs with throughholes for sensors, and valve etc. Swivel fittings were closer to $1.50 than the more typical $10-15 with eg. Bitspower. They look and feel identical to Bitspower - so I suspect these are "nightshift runs" of the same thing, B-grade sorted Bitspower fittings due to minor cosmetic defects, or really good knock-offs. I got one that was very soft and leaking, but all others were nice and tight, and at that price I did order spares. Also got a bunch of replacement O-rings as I wasn't 100% comfortable with the old seals on some of the 2nd hand parts. Also got 11.2mm drill bit, G1/4-19 threading tap and same size die, for making custom parts and fittings. Dirt cheap and very reasonable quality, just gotta educate yourself on what to look out for, and sifting through reviews to irk out the genuine ones. Compression fittings I got most of them with the other 2nd hand EK stuff, only needed 2 extra, so I stuck with the same EK ones for that. For being "EKWB quality", EK's fittings are actually fairly cheap. At least if you buy them from EKWB directly.
  3. I already wrote this up really nice and tidy, and had a power outage. This time will be quick n' dirty. I came across this cheap, nifty little gem for those of us who like informative mods and "tools" - and put it to work. It's a nifty little LED display with controller for 10K thermistor display. Temps in Celsius. It is really small and discreet, visible through your case window even in daylight, but no so intense as to be distracting in darkness. Can be used with any 10k thermistor probe, doesn't have to be an in-loop one. For what I did with it, you'll need: Mini LED display - $3 (Got mine here: AliExpress ) G1/4" Plug with 10k thermistor temperatire sensor - $2 (Got mine here: AliExpress ) Some scrap leads and heatshrink Molex / Sata / USB / whatever adapter you can splice into for +12V and GND (aything between 4-28VDC will do) Soldering iron and solder Something for cutting and cleaning up a square hole in a spare PCI slot cover - dremel, file, drill etc. Display looks like this: Black and red power leads are way too short to be of any use for this. Desolder them, solder on scrap leads of ample length for tidy cable routing. Solder or otherwise wire other ends to +12V and GND on a molex / sata adapter, or another suitable source. Cable length on the G1/4" thermal sensor plug is way to short - extend as you need, solder joints and heatshrink over solder joints. Install the plug in suitable location in your water loop, using a free spot or a splitter/T block. Mark around the opening in the rear case panel, onto the PCI slot cover while it's installed. Cut ~23x10mm hole in the cover (Due to the design of my slot covers, I had no area to drill holes for screw mounting, so made the hole precise - friction fit for the display housing). Cut between the markings you made before, to make sure no part of the hole is covered up by the rear panel of the case when installing the bracket. Cut another hole, or a slot out towards the endge, in the slot cover for pulling the cables back into the case. If you cut a slot all the way down, it might save you some hassle routing the cable for power, as those two cables are soldered to the display PCB, not socketed. In my case I opened the space between a couple of the original holes. Some cable sleeving can be good, but optional. Pull the cables through the slot, install slot cover, pop the display in the hole from the outside pointing in. Oh and I used a black marker to color the white sides of the display housing. Slot cover got a bit scuffed up while filing the edges of the square hole, if you're picky about that, cover it in painters tape while working on it. Display can be had in other colors than red. Different temperature probes exist, any 10k thermistor will do (one is included with the display, fine for air temps. Listing says it's waterproof, I haven't tested that). Due to the small size and cost, doubling up on these and having one pre, one post the waterblocks could be nifty. Feel free to bring any questions, ideas and suggestions.
  4. Input latency on most reasonably priced TVs is kinda bollocks though. And if I understood your sentiment correctly, that kinda suggests that whatever it is you're playing, doesn't have good native ultrawide support. But many games do. Ever driven a car in RL? Your field of vision out of the car, is much more akin to ultrawide formats, than 16:9.
  5. Any plans for there being a NAW show archive at all, on Floatplane? Linus wasn't wrong on yesterdays show, about the audio quality (Floatplane >> YouTube). Unless I'm missing something, it does make sense to be able to enjoy it in all its floaty, planey goodness. And I'm pretty sure I'm not the only one that rarely get to catch the live stream; it runs at like 3.30AM here, and generally between 2-4/5AM across Europe.
  6. Can't imagine I'd ever want to leave Open Bionics if I ever got the chance to work there. Its very much the dream job, working with development, prototyping and outreach in that field. Open Bionics is kinda a special dream job for me, as its almost scary how much of a mirror image of the business initiative I always Imagined I'd establish myself one day, if no one else did by the time I finished my re-education.
  7. Far less bulk material in the housing to absorb sound with a 2.5", not only due to the miniaturized nature, but also as weight-saving is a priority for laptop (2.5") drives. They'd carve 3.5" drive housings out of slabs of osmium and uranium if it helped performance the other density.
  8. Ah yes. You're gonna want to reformat the SSD as GPT for UEFI compatability. Unless you need it, I would advice against running in legacy mode just to avoid formating the SSD to keep it as MBR (which I don't imagine will pose a problem as you're reinstalling your OS anyway).It can pose all kinds of headaches down the road, and the only way of changing it is reformat and reinstall OS. Lookie here - it gives instructions on how to boot as UEFI (not legacy), with a Windows 10 DVD or USB bootable drive/stick/device/memorycard/potato, and perform the format to GPT (if you're getting this error I expect you have a bootable install available at hand). About 2/3s down the page, look for sub-headline "To manually wipe a drive and convert it to GPT"
  9. No real difference. Just a different SKU for (different) marketing reasons.
  10. What drive, and how was it set up in the old system?
  11. 860 EVO drives don't come in 120/128 GB variant, nor are they that cheap. That said, if you can sqeeze room for it, they are good value. As noxdeouroboros suggested; wait for the heavily discounted offers right around the corner (Black Friday, more like Black Week), and see if you can squeeze in a 240/250 for around the same price as those you listed. I've built 4 budget systems the last year or two (Mediacenter NUCs, laptops, 2nd hand budget gaming), 3 of them with 120GB SSDs, and ended up regretting not going with 240/250 each and every time I chose 120. Also note; 120/128G SSDs are usually much lower performance tham the same range 240/250 and 480/500 model(s) - so it's not just a matter of size. Most retailers seem to be gearing up to battle over SSDs this year so there will be some solid deals to be had for sure. Already seen the A400 240 GB for $39 and A400 480 GB for $58, and "the week" hasn't even started yet.
  12. Personally, seeing as there is no VRM heatsink at all on that board and you're planning a "permanent" overclock, I would probably feel more comfortable in the long run with sticking either one heatsink cut to fit, or a couple of smaller ones, on the MOSFETs, and see to it that there is actually a bit of air passing over them. I'm thinking generic heatsinks for smaller components here, not something custom (beyond cutting one to fit perfectly if you want to go that route). Just google "vram heatsink" and you'll understand what I mean if unclear. These can be had in pretty much any shape and size for dirt cheap, and lots of different colors, so finding some that go with your asthetics without needing to paint them, is rarely a problem. I would probably get some 20-30mm tall, if there is no space conflict with CPU cooler etc. Some of them do come with thermal adhesive tape-pads, although quality and adhesive properties of those can be a bit shoddy for the very cheapest heatsinks. For this usecase, I would personally consider using a proper thermal adhesive though (although that does mean they won't be coming off again). And try to make sure that if you go with a heatsink with long fins instead of "individual bars", that the fins are oriented vertically when in place, not horizontal. Heat rising and all that. This might not stricly be absolutely neccessary, but given how little it costs in both effort and dineros, I would personally feel alot more comfortable with some simple generic heatsinks on those MOSFETs.
  13. Regarding LM between IHS and die, and IHS and coldplate / cooler base, is the risk of LM reaching beyond the area of application. The risk of spills are much lower between IHS and die, as that would strongly suggest that you're putting the IHS back on. Thus pretty much making it an almost-sealed chamber, even if you don't glue it back on and its retained "only" by the force of the retention bracket (and/or mounting force of cooler). If applying superglue, nail polish etc over the SMDs under the IHS, any minor LM spillage won't do damage as long as it is contained "under" the IHS. Between the IHS and the coldplate however, there are plenty of mischievous places for the LM to go. Beyond that, as with GPUs, there is significantly less to gain from using LM over a good quality non-LM TIM, due to the much larger thermal transfer area available. Another worthy of note tidbit; Many SMD components, especially larger ones like shunt resistors on graphics cards, contain aluminium, and tin, in their end-caps. That means Gallistan and other LM TIM alloys will not only deteriorate the solder, but the contacting cap of the component themselves. So using LM for "shorting shunts" is really a horrible idea. There are plenty of other options that do the same job wihout soldering, like rear window defogger repair "paint" for autos.
  14. This could be a sign of (somewhat early) permiation, depending on its age. It looks like its either a H100i v2 or a H100i GTX. Could be as old as 3 years. But first try a few quick tricks and see if any experienced changes "stick": As airdeano suggested, shim the radiator so that it is at a sliiight angle, the rear (opposite the hose connectors) end being the highest point. Any air trapped should accumulate there, and mostly stay out of loop circulation. I would also suggest tilting the case so that the pump/block assembly is the lowest point in the loop including the hose bends, then turn either only the AIO, or the computer, on and off several times while keeping the case in this position. Prop it up securely while doing this, or get a helping hand. Then set it back down. Report back with results.
  15. There are plenty of alternative options for improving VRM cooling besides the blocks included with the Fomula. Anything from ensuring decent airflow to the stock heatsink, to custom / adjustable 3rd party VRM water blocks. All of which will get the job done, well, at a far lower price tag than the price gap between Hero and Formula. It all depends on your preferences, wants and needs. If the Hero covers these except for VRM cooling, I would suggest going with that, and start with airflow - that will assist in cooling other nearby components aswell. Procure (unless you already have (access to)) a thermocouple or other suitable measuring equipment, and observe actual VRM temperature during your overclocking adventures. Unless the temperatures actually turn out to go worryingly high, there would be no need to progress the cooling ladder further. The several hundred dollar gap between Hero and Formula would in this case be much better spent elsewhere. Don't know about Suomi prices, but here in Norway the price difference between Hero and Formula isn't far from the gap between a 1070 and a 1080ti. There is always the "because I can" / "because I want to" element - but thats a whole different matter, and really not subject to the same criteria.