Mr. Fox

Yes, he sure did.

Cool, I thought that might be the case. I just tuned this cheap G.SKILL 7200 kit to 8000 stable last night. I can't tighten the timings futher because it gets INSANELY hot with the crappy stock heating blankets. But, pushing 80°C in a TM5 test without errors is definitely stable. The low CPU voltage requirements on this chip is nice, but still a bit much for a 360 AIO to manage. Delid will help. It will be interesting to see how this CPU runs bare die in one of the Apex systems with much better thermals. @tps3443 here is what I managed in spite of the scortching temps. This is with a crazy loud Noctua fan blowing a ton of high-speed air onto the RAM sticks. I am impressed that such a small mobo with only as many phases (only 10 phases) as they can cram onto this baby-sized PCB can do this without bursting into flames or exploding. @Rage Set maybe you can do the same with yours, bro.

What system are we talking about? Was it this ITX board?

Yeah, that really does suck and I am absolutely opposed to it. I don't have a problem with a software company (games or any other kind of software) making money. I want them to or they will go out of business. But, I don't want them to milk me for money after the sale. I view that as being dishonest and will generally choose to go with another solution or just do without and not buy anything from them. Games should be inexpensive. Making them free and charging for the add-ons needed to make the experience a good one is a dishonest shell game. I say figure out how much you need to charge to turn a profit giving all of the features and functions as standard, charge that for it and say take it or leave it. If I want it and it is priced right I will buy it. If not, I won't. And, if I know I am going to be milked to death I sure as hell won't. Selling DLC, skins, add-ons, etc. is a scummy way to make a sale that keeps on selling. Speaking of "milking" I finally made an exception for Office 365 exclusively, and for two reasons. I am renting 6TB of cloud storage for $99 a year and can't do that anywhere else that cheap, and the software is essentially free. Second reason is that it is a "user license" not a "machine license" so I can install as many times as I want on up to 6 systems simultaneously and de-register installations when the system is gone or I have done a clean OS install, so it is essentially an unlimited number of installations. The "pay once/own forever" lifetime Office license is a scam. You can only activate it a certain number of times on one machine and then your license is no bueno.

I do not disagree, but adhering to that rigidly I could not buy anything from AMD or NVIDIA. I view them both, along with many other large corporations (especially those in technology) as ethically bankrupt and fundamentally corrupt and dishonest organizations.

I think in some ways you understood my point and in other ways missed it. We are talking the same language but interpreting things with subtle twists simply because forum posts are often a poor substitute for conversation. Having a personal preference is normal. We all do, and if we deny it we are not being completely honest. The only thing I am critical about is the foolishness of hating on something based on brand. Brand should not matter. Results do and should matter, but they are relative to the intended use scenario. Taking a position that one doesn't like a particular brand because it is inferior in the metrics they care about or need the product to deliver is valid. Price adds a slightly different dynamic. One might choose the inferior product because the superior product is so grossly overpriced that its benefits are eclipsed due to a price so high it cannot be tolerated. Thus, settling for the inferior product might be the smartest alternative. But it is still inferior. If the brand one prefers doesn't perform well in the area they intend to use it, buying it based primarily on an irrational hatred of a brand is proof that the person buying it is an irrational person that lacks good judgment (a fanboy). The intent might be to punish the brand they dislike and reward the brand they prefer, but their purchase decision will accomplish neither of those objectives. I am relatively confident that Intel, AMD and NVIDIA don't care what we want as individuals. They only care about what (and how much) they can sell us collectively. That's very accurate. I have changed my approach in the past year or so and I think the smartest move is to wait, watch and buy used when something newer displaces it whenever possible. That's easier to do with things like motherboards and CPUs than it is GPUs. GPUs are more prone to failure and good samples are not sold and replaced by an upgraded product as quickly (or so it seems). It is better to grab a silicon lottery winner second hand at used pricing from someone with the "latest and greatest sickness" driving their decisions whenver possible. Although there can be legitimate reasons for it, such as unexpected financial hardship, it is generally a red flag when a current generation flagship GPU is being offered for sale. There is a good indication something is wrong with it. A silicon lottery loser or some other factor driving regret and dissatisfaction on the part of the person selling it. If it is a truly "special" sample it will generally be offered up at a scalper price because the seller sees the opportunity to profit and they value that opportunity more than the product they are selling, or the person they are taking advantage of. The golden sample 13900KS I just purchased for a price that resembles MSRP would or could have been sold for double MSRP before the 14900K appeared. In fact, the person I bought it from was selling it to offset the cost of his CPU upgrade; not because he wanted to scalp someone feeling compelled to have a better current generation CPU at an insane price. It is a reversal of positions, but it is smarter to take advantage of the opportunity to buy at the correct price from a person needing money to offset the cost of the shiny objects they are chasing versus buying high from a scalper that doesn't care about the shiny objects.

There will be some business case scenarios where that matters most, but those are generally not selling points for anyone buying a GPU for gaming. Poor ray tracing performance is a real problem even though the underdog fanboys dismiss it. It wasn't very important back in the Turing days but, like it or not, that is important now. And, I am sold on how much better the image quality is with ray tracing and path tracing. For content creator purposes Red Team doesn't bring much to the table, while CUDA gives an edge to the Green Goblin and GDDR6X is faster. So, it depends on what you are comparing it to. Having more VRAM is important when you don't have enough. When you do have enough (or things are equal) then the faster memory is superior and becomes a no-brainer on face value. If you look hard enough you can find at least one area where one brand is better than the other, but you shouldn't have to look hard to identify why or under what circumstances it matters.

Same. Superior is superior. I would actually love to see that happen. If a person is going to spend $1,000+ on a part, having it take 2-3 years to become obselete and replaced by something better is a win at the consumer level. Not going to make highly paid execs and shareholders happy, but nothing will please them other than more money so their happiness doesn't need to matter to me/us. They can't compete at the high end on performance or features, so the only area left for them to try to compete is price. Because their performance and features are second fiddle they really do need to compete on price. Otherwise, it is only the fanboys and underdog-cheering group that find them compelling. Finding them compelling for no reason other than because they are "not NVIDIA" and "not Intel" is irrational and has no merit and it's hard to respect. Finding them compelling because they are the cheapest option is a totally legitimate reason if that is the buyer's prime objective and it is easy to respect that rationale.

I know some people like the FE models but that would be my last choice unless or until they start making them with dual vBIOS. Personally, I would go with the AIRO Extreme or Suprim. AIRO is the only one left in stock at NewEgg right now that I would go for. Nothing left at the local Best Buy. Glad I am not shopping for one. If I were, then the AIRO would be the only good choice for me. https://www.newegg.com/zotac-geforce-rtx-4080-super-zt-d40820b-10p/p/N82E16814500579 https://www.newegg.com/msi-geforce-rtx-4080-super-16g-suprim-x/p/N82E16814137853 Or, a TUF would be OK But, they're all gone as well. https://www.newegg.com/asus-geforce-rtx-4080-super-tuf-rtx4080s-o16g-gaming/p/N82E16814126693 Hilarious to see ASUS still scamming their ROG fanboys with the ugly-looking, oversized, overpriced and underperforming Strix abortion GPU. The one thing that would make me lean toward steering clear of any and all ASUS GPUs is the upside-down 12VHPWR connector that makes them a morphodite freak. Proprietary oddities like that are not something that makes good sense. You can't even use the cable provided with an ATX 5.0 PSU if you lack the judgment of avoiding an ASUS GPU.

That looks very nice. Tasteful decorum. Our electric bill here is about $125 in the winter (seldom need heat) and around $375-$400 due to HVAC overload when mother nature's blast furnace is running. Edit: OK... minor changes to primary timings and tREFI, but substantial improvement on latency.

If Battlemage hits the shopping cart at around $500 and falls within slapping distance of 7900 XTX/4070 Ti/4080 it will be a financial upset to AMD and NVIDIA. Apart from that, I think the only thing we can count on with GPUs is getting screwed over on pricing. This 13900KS seems really good. I still have it in the ITX system with that same sheet of indium instead of thermal paste. I have all P-cores at 57x, E-cores at 47x, Cache at 50x with 1.250V and -0.075V offset and MSI LLC 4. VID and VCore are almost exactly the same value. Using a sheet of indium has never worked well for me. I bought the material as an experiment. It never melts. I peeled it off the Celeron and layed it on the 13900KS and reassembled, expecting tragic thermals. But, in spite of that the temperatures are not nearly as horrible as I expected them to be. I've ordered a cheap A-die kit to replace this Crucial DDR5 Pro kit I got for free. I am running it at 6000 with sloppy timings and it's not that great. Nor did I expect it to be. Latency with a capital L with this mainstream memory kit. Tastes and smells like a Ryzen latency, maybe even a little worse, with this mainstream zombie memory, LOL. To be fair, all I did was change the memory speed to 6000 and bumped the voltage from 1.100V to 1.300V, did not tune anything on the memory. It booted and ran stable, so there is room for improvement (although maybe not worth the effort). I have no idea how to interpret the MSI CPU Force rating. I have not really started tuning anything since I am using this as my work PC right now and I've spent most of my time on ASUS software cancer exorcism (wow, what a chore). I just moved the NVMe image from the Encore to the MSI board. So far this morning, I am just randomly setting BIOS values (everything was on "Auto" before) to see what happens, so I am fairly confident I can go lower with the voltage and the temps would certainly drop with some KPX or MX6 in place of the indium sheet.

I should sell my 4090 and use the money to buy two 4080S and a 14900K. It is sad that the only product that is worthy of respect is priced in a manner that is deserving ridicule and condemnation. To be more precise, the people responsible for those circumstances are deserving of ridicule and condemnation. They are unworthy of gainful employment, a secure future, happiness and good health. Fortunes that last are built upon generosity, not greed and self-serving interests, are the only kind that will stand the test of time.

Well, still no CPU. USPS keeps "parking" it at different places for several days for no reason that I can identify. It sat in NY for several days, then moved to the Phoenix hub and sat there for three days, and now it is in transit to my local post office (finally). It looks as though I might have it tomorrow. So, for now I finished the build using a dual core Celeron from my spare parts stash as a temporary placeholder. Cable management wasn't as difficult as I expected, but this isn't a ridiculously small chassis (thank goodness). I am using a sheet of indium cut to the size of the IHS and the Celeron stays under 40°C with a load test, LOL. That will leave me with no thermal paste to clean up when the 13900KS from Sweden is delivered. I will use it in the midget PC for a while then delid it and move it to one of the Apex machines to run bare die, then put the lowest SP quality 13900KS into the midget PC. This is the motherboard I purchased from Brother @electrosoft. It finally arrived today. Not sure what the deal is with USPS, but they're really messed up right now. They suck more than I have ever seen them suck before. Our tax dollars hard at work. Thanks, Brandon. I've got more messes to clean up with the big Dark Base Pro on a wheeled platform down below. Will have to figure out how to manage the extra stuff next weekend, but everything is connected and working in spite of the wiring and tubing rat's nest. Just for giggles... nice temps from both cores, LOL. Thank you. It was a pleasant but uneventful day. I hope your birthday was a good one. too.

Thanks, brothers. I really appreciate it. Hopefully the motherboard from @electrosoft and the SP117 13900KS from Sweden will arrive on Monday. I did find a way to mount the 4x 1TB SATA SSDs on the PSU bracket just above the GPU. Plenty of room for a long GPU, 360 AIO and normal ATX PSU. The A770 has several inches to spare. Should have plenty of airflow with 3x intake and 3x exhaust.

Yes sir. Thanks.

Just crested beyond the official half-way mark today. Some say that's over the hill, but by my calculations it is still a long way to the top.

No. Just. No. Only an idiot would think that is a good idea. But, OMG, we are surrounded and outnumbered by idiots. So, I would not have a surprised look on my face. I expect 9 out of 10 things that happen in tech world to be stupid due to how prevalent idiocy has become. That applies to producers and consumers. Both are extremely unimpressive groups dominated by a majority of degenerates.

In some ways it probably bears a resemblance to the satisfaction we once had in laptops with desktop CPUs and high powered GPUs, dual 330W AC adapters, etc. They never worked as well as a desktop, but they highlighted how sucky pansy skank turdbooks are.

And, what overclocker wants to get screwed out of enjoying memory overclocking by being cursed with a 4-DIMM motherboard?

Almost anything you buy now is going to be hobbled and castrated. If I were going to buy a brand new GPU right now and my budget was $1200 or less it would be only 4080S or save $900 for later and go with an A770. I would not buy anything from AMD or accept anything less than a 4080S from NVIDIA. If you want something fun to play with, the A770 core overclocks better than the core overclocking on anything sold by AMD or NVIDIA. As long as you don't need the horsepower of a 4080S or 4090 it's the best bang for the buck. The memory does not overclock (no way to adjust it) but the stock memory clock on my A770 is within 150-200 MHz of the max stable overclock that my 6900 XT was capable of handling and it doesn't drop clocks under stress like the 6900 XT did. I would take that budget 2-DIMM MSI mobo over any 4-DIMM mobo without even pausing for a moment to think about it. 4-DIMM motherboard are a waste of money for overclocking or gaming. Not to mention that they all suck at memory overclocking.

^^^ A year ago... ^^^ Now...

Wait. Spend the $200 for the better GPU. My 2¢.

It fails against an MSI Z790i motherboard that clocks the RAM at 8000 because it has 2 DIMM slots. Yes, that's right.... a $300 MSI mini-ITX that kicks the grossly overpriced $700 ROG turd's digital butt. There was exactly such a company. Until about a year ago. They made the world's best motherboards and GPUs. Now they only make stupid crap like power supplies, mice, keyboards and video capture junk. Gee... thanks, NVIDIA.

It's both sides being stupid. With ASUS it is the combination of being dishonest and stupid. They do something right once in a while but screw things up just as often, if not more often. For every good product they produce, there are at least a dozen products that are pieces of crap. Their successes are inconsistent and unpredictable.

Having done this for many years I do not recall having ever perma-bricked a GPU (laptop or desktop) flashing the vBIOS. Probably have done this more than 1,000 times without incident. Yes, it can happen if you are careless, but even if it does you can easily fix it if you know how. The brick is not permanent. You can blind flash if there is no display output. You can use a manual SPI flash programmer in the worst case scenario of a hard-brick. Inconvenient but not irreversible. You can save the vBIOS using GPU-Z or NVFLASH. I prefer using NVFLASH. nvflash64.exe -b filename.rom This will save the firmware dump in the folder where nvflash resides using the name you provide in the command syntax. Although rare, I have had motherboard BIOS updates using an "official" firmware released and recommended by the OEM go south a couple of times. Compared to 0% for GPUs, flashing a GPU is much safer. Also, if you identify an error flashing a GPU, if you keep the system running and reflash to original firmware without shutting down you're generally good to go because the new firmware (or corrupted flash) doesn't take effect until you reboot. When working with SVL7 and Prema on firmware mods there were times when I flashed the same GPU more than 100 times until the mod got dialed in where we wanted it. With modern GPUs it is not possible to flash unsigned firmware using NVFLASH, and if you force flash it using an SPI programmer sometimes it will not function correctly, depending on what was modified. As long as you take a dump of the original vBIOS chip, the fix is generally as simple as flashing it back. I've never perma-bricked a GPU using an SPI programmer. You can also use an SPI flash programmer to fix a motherboard flash that went wrong. This is generally only necessary with a GPU when you are experimenting with mods. Never had any need for it cross-flashing official signed firmware from a different GPU in the same model line with a higher power limit. Edit: One advantage to having a dual (or triple) BIOS or vBIOS is if something does somehow go wrong you can slide the switch to the other position, power up and once in Windows, slide the switch back to the other position that you screwed things up and flash it. This will work with all GPUs and all motherboards that have a legitimate hardware toggle switch. Sadly, some of the most expensive enthusiast motherboards (including the Apex and higher ROG mobos) DO NOT have the right kind of BIOS switching features on the motherboard and that tremendous benefit is lost. You can't switch to the other firmware position after the motherboard is powered on, which is absolutely absurd, sinful and deplorable on the part of the psychotic control freak manufacturers. Never buy an expensive GPU that doesn't offer a dual vBIOS. Not very smart on the part of the manufacturer or the consumer. There is no excuse for it to be made that way unless it is a cheap low-end piece of garbage. Certainly inexcusable and unforgivable on an FE card.