electrosoft

M5 Ultra vs a 9950X3D + 5090 kitted out PC: CPU: Bonkers, next level CPU performance. Absolutely trashes Intel and AMD at 150w with their 36 core model both single and multi. Single core: ~44% faster Multi core: ~80% faster All this while consuming 25% less power in a small, compact box that makes zero noise 99% of the time and even under load is a fraction of the noise a typical PC generates. You're going to need a Threadripper 9970x (32 core) to compete, but not beat, at this point on multi. To flat out beat Apple, you will need a 9980x. On both, they absolutely get destroyed for single core performance. On Single, Intel and AMD are going to have to bring some serious magic in their next gen CPUs or Apple will just continue to dominate and I don't think they're going to even get close to catch Apple in single threaded performance. Looking forward to the 52 core Intel Nova "Open all salvos!" Lake beast vs the M5 Ultra. Heck, even the 28 core might get close in multi. Blender Render = M5 Ultra is ~72% faster GPU: Time to come back to reality Apple vs Nvidia..... Optimized for Nvidia, 5090 pulls ahead showing Nvidia's strength in terms of raw power, vertical integration and software stacks. LLMs: If memory isn't an issue, 5090 is a beast but obviously 32GB can't stack up against 96GB or 256GB. Toss in an RTX 6000 (or two or three) and it's game over but then again you can score a fully kitted M5 Ultra with 96GB of memory and 1TB storage and max CPU/GPU config for $6799.99 for less than half the price of an RTX 6000. You can score a 256GB model for $10799 which is 2/3 the cost of a single RTX 6000 at this point. In the end, Apple's massively upgraded Neural Processors are no match for Nvidia's raw power but their unified memory architecture is where they have always shined. Expect the 512GB model (not available yet) to cost ~$17k as the 256GB model is currently $10,799.,99 and upgrading from 96GB to 256GB is a $4k up charge. So you end up with a 512GB M5 Ultra for what is quickly becoming the cost of a single RTX 6000. Video editing, M5 Ultra: DaVinci = tied Abobe Pro = M5 Ultra finished in less than half the time Gaming: Time for a brutal reality check. Even Metal optimized, 5090 just trashes the M5 Ultra and is over twice as fast. M5 Ultra looks to be right between 5070ti-5080 in gaming performance but looking at some of the results that are actually taking advantage of the 80 core GPU, there is some optimizations to be had but....yeah, the 5090 just drops the hammer on Apple....just like Nvidia dropped the hammer on AMD with the 5090. My final verdict? M5 CPU architecture is next level. Absolutely next level. GPU has really come along way even vs M4 and its strength lies in its unified memory bandwidth especially in these AI driven days. For my personal use, target LLMs for particular use cases the 5090's 32GB is working out just fine and this is where it absolutely shines. If you need more than 32GB, performance scarily drops off as expected and you need to start looking at the 5080 based 4500 card with 48GB or even 5500 (barf) up to a 6000 to match the 96GB config from Apple. 256GB config for $10,799 really makes a compelling config for many small AI shops, startups or individuals because it costs less than a single 6000 by almost $5k and you get over 2.5x the memory. I suspect the 512GB will also be a very compelling buy for those on a budget but needing 512GB of memory.

Yeah, the 5080 really has the headroom too fully appreciate overclocking at what tops out at an extreme but "sane"ish type level vs the 5090 requiring a nuclear reactor. Extrapolating, that Steel Nomad is like scoring ~ 21k+ on a 5090... That Timespy score is bonkers both GPU, CPU and combined... I see you right behind Jwick on the bot in the 2 slot a few times, lots of top 5/10 finishes and #1 in TSE! Excellent work!

Holy crap..... 825w through a 5080?! Sheesh.

Technically, it comes down to what resolution and detail levels you want to run and expected fps and then apply DLSS5 over it. If you're running 1080p medium you can drastically adjust down to even a 5060 paired with a 5050. If you're wanting to run 4k ultra the same rules apply and depending on game, a 5090 paired with a 5070ti would be optimal, but looking at that WoW video, it is paired with a 5070 and handling it ok too. So far, it looks like when paired properly, you achieve close to your original fps with DLSS5 enabled using a second card to handle DLSS5 only. DLSS5 is known to up the power draw overall (there have been some reports of burning cards trying to run DLSS5 on the same GPU) but what else is new? But if you're off rendering DLSS5 that would, in the case of the 5090 at least, leave that extra headroom for a standard 600w card to extract maximum performance and then have your lesser card render DLSS5 only. I haven't come across someone running dual monitoring or a HWinfo log to see how much is actually pulled. The beautiful thing is with the current wrapper, even 9070xt supports DLSS5. As always, it will trip anti cheats so you either have to play single player games only or as Jansn did, pass through the output externally for processing and display. How is GPU utilization overall running QHD with the 5090m and 9955HX3D? So many think WoW could run on a tin can and it definitely does not. So many times zipping around in WoW at 4k ultra RT enabled and I routinely see the 5090 hitting 90%+ utilization which lets me know a 5080, like the 9070xt, would be sitting capped out more often than not at lower fps. I even see some utilization spikes into the 80s even in the city with the 9800X3D tuned up. Since those initial runs I posted for your comparison to your monster MSI with the same CPU, I haven't gotten back to the G16 yet but I'll get back over there for more tuning and gaming sessions. I have 6 days to return it, but I'm definitely keeping it for several reasons. First and foremost is how quiet it runs and the only way I can get it to get the fans going is to put it under meaningful stress. Two, I'm not hating the smaller PSU and much smaller and lighter chassis. It reminds me of the Clevo NH55 12900k + 3070ti in size and weight. I really like the 240hz display. It isn't as bright and in your face as the Alienware, but much more balanced. It is flat out beating my Alienware with the 275HX in gaming in WoW and especially in Fallout 76 in my test spots. Lastly, cost.

World of Warcraft current retail Midnight DLSS5 using a 5090 primary and DLSS5 external rendering with a 5070.... Talking to Oglethorpe over on OCN, apparently the 5090 pairs best with a 5070ti and a 5080 with a 5070. 5090 + 5080 is slight overkill but that's ok. In current market conditions, 5080 + 5070 is going to be the sweet spot methinks.... Once we get around the anti cheat problems and external DLSS5 rendering refinements, this is going to be the god mode way of playing.... Same goes for Fallout 76. Same anti cheat problems atm, but here is Fallout 4 (same engine and basically graphics) with DLSS5 enabled. Absolutely gorgeous. More realistic (IE not modded to the gills) Fallout 4 with DLSS5 is still crazy good looking: I said it before, but I'll said it again. DLSS5 is such a more radical leap forward for graphics fidelity than Ray Tracing it isn't even in the same galaxy.

Not that we're surprised and it was fully expected, but possibly no new GPUs from AMD and Nvidia till 2028..... If the market demand keeps up as is, I wouldn't be surprised to see a fall 2028 release and an almost full 3yr cycle vs 2yr. https://videocardz.com/newz/nvidia-geforce-rubin-and-most-amd-rdna-5-gpus-reportedly-not-coming-until-2028 ---------------------------------- Another nifty tool that raises the power limits of Nvidia laptop GPU: https://videocardz.com/newz/modder-creates-tool-that-raises-rtx-5070-ti-laptop-gpu-power-to-180w-rtx-5090-up-to-225w Not sure what we will see, but I know my OC'd Alienware 5070ti at 175w wasn't much, if at all, over my G16 9955HX3D 5070ti OC 140w. @win32asmguy push that 5090 to 225w..... ----------------------------------- Might need to put a bit of clear tape over your serial # moving.....the serial # isn't stamped somewhere on the PCB too?? Regardless, common Nvidia..... -------------------------------------------------------------------- DeepSeek is the answer to preventing total AI dominance by a single company or a collection thereof? https://wccftech.com/deepseek-engineer-stuck-in-self-destructive-loop-says-building-ai-to-replace-him-is-the-only-way-to-stop-openai-anthropic-from-creating-a-cyberpunk-2077-society/

Checkmate on the consumer grade 5090 availability. 😞 Now at least we know why 5090s have been drying up AND prices going up. If AI shops are converting 5090s to AI farm cards as is and with extra memory, why not just re-direct those 5090s into RTX 5500's with 84GB of memory which is close enough to the RTX 6000 with 96GB of memory and charge $12k? Scary thing is if demand doesn't abate, all pricing will continue to scale upward.. Demand is so high, and memory capacity so important, that they are now scavenging their top end consumer grade GPUs..... We already have the RTX 4500 using the 5080 (GB203) die with 32GB of memory for $4500. If Nvidia decides to bump that to 48GB+, we could see a sudden shifting of even those dies at this rate. Just wow..... That is a hard NO. I mean zero chance. Zilch. Nada. Nunca. Nyet. I wouldn't even pay $4k let alone $7-8k for a 6090. I can't begin to tell you how hard it was to swallow that $3399 I paid for the Astral which I LOVED and returned purely based on price alone. Snagging a Vanguard for sub $2k and a Ventus 5090 for $2k last August was lightning in a bottle..... I'm kicking myself for not keeping the open box Ventus 5090 for $2k from last August 2025..... 🤣 You might be kicking yourself a little too for not holding onto the 5090 FE considering how much the net now. But everyone has a breaking point and if 6090s are $3.5k+ at launch which is A> Extremely conservative and B> Assuming we even get 6090s seeing as how the 5090s are being redirected into AI sector level cards due to demand, who knows? Who even knows if we get a 2027 launch as 2028 is looking more realistic unless AMD brings the high heat (which rumors are still saying YES). It really feels like Nvidia is trending pricing up and next gen the 6080 will slot in right where the 3090 used to be price wise at $1499.99 and it will be basically our top end card next go around. 4080 Super is basically on the heels of the 5080 so a performance bump for the 6080 to maybe 5090 next time or at least between the 4090 and 5090? More AI tech as it refines to continue to boost performance and visuals while working on latency. Meanwhile, reserve the top end dies for the RTX 6500/7000 class cards for AI. But yeah, when we get to the point of $5k+ 6090s at (paper) launch and memory kits hitting $1k+ for 2x16GB let alone CPUs and MBs combos of meaningful SKUs hitting $1.3k+ it is time to downsize and look to longevity of current hardware if that. I can see a time where I sell off everything, quietly step down to 080 (or lower) tier hardware and mid tier CPUs and GPUs and work with that with these crazy climbing prices. Once you adjust your enthusiast POV, you suddenly find yourself building out much cheaper systems that are quite capable. I'm to that point I won't even spend more than $3k on a PC laptop tops and even there I tend to trend lower than higher. I had the opportunity to pick up an MSI Raider like @win32asmguy has but it was $3650 + tax and I walked away especially in light of the $1800 for this G16 which takes a major performance hit (~18k TS vs ~26k TS) but is holding it down at half the cost with a 9955HX3D inside. My costs incurred this year for problems and issues is mind boggling from incidentals to a financial and working stake in a small start up. The hits just keep on coming. 🤪 Quite honestly, the smart move for me is to sell off almost everything and pick up a Studio M5 Ultra w/ 96GB as the 48GB in my (albeit contractor issued) current M4 Max is hitting a wall for the small startup I'm involved with as a side project atm. Whoa! Very nice and a good price in these trying times! I'm glad to know that site is legit and I'll be adding it to my list of places to check out on the regular, good find! Do you plan on putting better heatsinks / water cooling any of your sticks? You already have a good loop setup so it would be fairly straight forward to get them running uber cool. Our local MCs still have these in stock along with a few 5090s namely the GB Infinity for $5400 and I saw an Astral 5090 AIO for $5229. I just checked and there's a couple white Astrals for $5k+ but I also saw a 5090 TUF for $4699.....because that's kinda a "deal" atm? Ugh.... MC sells the Asus bundle for $10829 = ~$11.5k after tax. That. Is. Insane.

This is the growing theory and something we discuss in a few think tank groups. Right now, Nvidia has such tight vertical integration they rule the roost. They also lead in performance but the integration and software is key. Rest assured, AI focused companies like OpenAI, Claude, Anthropic, etc... are working towards common stacks to integrate ASIC for hyperscalers to get off of Nvidia's teet along with smaller operators. On the other hand, you can see the writing on the wall with China and them slowly, but surely, moving everything in house or how DeepSeek took everyone by surprise showing a potential future of better optimizations and larger, wider scaling. Right now, Nvidia is IBM circa 1975-80ish but it won't last forever. AMD is working feverishly on their next gen AI products. Nvidia is entering the CPU foray because they realize eventually CPUs will increase in importance for AI performance in the long run. We're still in the wild wild west phase atm, but the moment normalization/common models are established and/or open standards, Nvidia is in trouble. Throw in some ASIC level computational processing? They're in REAL trouble. One thing that won't change is the need for high bandwidth memory but a sudden and abrupt shift in needing expensive GPUs to ASIC like devices like what happened with crypto mining would be beautiful to behold. Right now, I wish the 5070ti (even some 5080) and down potential customer base would all switch to AMD for the cost savings alone and when Nvidia looks up, they've lost major market share going forward for when that day comes but as we saw with the last market reports, Nvidia actually gained back half the market share they lost to AMD in the consumer gaming sector.....crazy.

WoW with DLSS5. Big frame hit, but hopefully they sort out a solution up to and including adding in an additional GPU to off render on the secondary card. Visually much more impactful than RT ever was IMHO Absolutely gorgeous and yes I would run this in a heartbeat when playable... Maybe see about slapping the 5080 in to run along side the 5090 if required. ------------------------------------------------------------------------------------- Somebody snagged it already lol. Did your other kit from that other site ship yet? I do remember him vividly along with his multiple bannings. Sorry to hear of his passing. AMD chips are sensitive odd ducks in that I've found more often than Intel chips that too much voltage can cause more instabilities and lowering definitely helps including macro voltages like VSOC to micro voltages like VDDP. This is regardless of thermals where you can have them locked down and those overvoltages can be just as destabilizing as undervoltages. It does make me wonder how your two follow up chips would have run on the Dark Hero but then again it could have been a dud board too considering it was a return. Glad to see #2 and the 2200 has the stronger cores. Still #3 is the clear winner for 1:1 overall but if #2 rules the roost at 8000+ I would still go with that one. I know BB sent you two, but did they note on the invoice/receipt the serial # of the one meant to be actually sent to see which one you could actually return? Worse case scenario, you can just sell off the second one though.

Ouch, well as I've found over the years, wife requests override personal projects. 🙂 I didn't even know they "refreshed" the 2025 SKU with a 2026 SKU so I went and took a look. So basically, a 240hz->300hz display and jacking the price up from the original MSRP of $2499.99 -> $3299.99....yikes. Yeah, mine was open box for ~$1800 (1805.28 to be exact) listed as "used, acceptable" but arrived looking brand new with all accessories. Only the box was a little beat up. I realized I hadn't done anything with it and the return window is less than two weeks away (22nd) so I went ahead and finished setting it up and ran some basic benchmarks before I start tuning with the BIOS OC controls. Timespy performance is basically equal to Alienware performance mode just the G16 pulls 35w less. 9955HX3D definitely beats my Alienware 275hx in performance mode CPU wise in CB23. I said it before and I'll say it again, Asus fans have some of the best pitch I've heard on laptop fans. Definitely front and present, but max for max, I've always preferred their overall acoustics. Reviews were right. Even in Performance mode, this thing is nearly dead silent doing everything not stressful and even then the fans didn't really kick up that much. It took Turbo to bring it to life and Manual to really let you know it can move some air. All tests were run on a flat, hard surface with zero rear elevation which is how I would normally run it on the go with some laptop balls. I'll end up spending tomorrow playing and comparing Fallout 76 and WoW with my Alienware data but I expect the 9955HX3D to pretty much own it badly in Fallout 76 and beat it overall in WoW because of the lack of tuning available for the 275HX and dog slow 6400 memory. The real performance question will be can the 9955HX3D cap out the 5070ti in QHD in Fallout 76 because the 275HX couldn't only in pockets so we shall see. If it can really cap out the 5070ti, that makes a compelling argument to get a 9955HX3D + 5090. If it doesn't, then the G16 it is. Some data to compare against your Raider with the 5090 before I dive into the BIOS OC options.... Performance (default) mode: Turbo Mode: Manual Mode (All sliders to Max): ------------------------------------------------------ Nice snag! I was away from the forums over there so I just checked and saw the poor arguments for the price being asked. If they think prices are bad NOW, wait till next year..... I'm always on the lookout for good sets to grab, and I would have grabbed those and jettisoned my KingBank's to eBay or similar (or sell them in a parts constructed system). Actually, those KB sticks in the 265k box did 8933 so I shouldn't poo poo on them too bad. It's just their poor AsRock and Asus fails that has me salty. 🙂

AMD 9070xt officially AMD's most popular GPU on Steam....at least some buyers understand it is the best bang:buck on the market by far --------------------------------------------------------------------------------------------------- Newegg is sold out of all 5090s (shipped and sold) except a combo deal with the Asus 20th edition combo along with a 9950X3D2 Amazon shipped and sold out of stock. Just scalpers scalping as always..... Best Buy is completely sold out of 5090s BH Photo is completely out of stock My local Microcenter only has three 5090s in stock with a grand total of 4 units on hand all $5k+ and the Gigabyte Infinity has gone up $100: @Papusan how is the stock situation over in the EU? ----------------------------------------------------------------------------------- Man.... if you had just held onto it a bit longer you could have gotten 50% more for it.... That lack of BIOS flashing was the deal breaker.... Astral is still the best looking card this generation IMHO but I do love the lit up Dragons on my Vanguard. While you did pay a bit more over an already inflated MSRP, the fact it goes for $5k now if you can even find it in stock was a good call if you wanted to ride the 5090 to 2028. -------------------------------------------------------------------- @win32asmguy DLSS5 on WoW looks pretty bonkers graphics wise, wow..... Any updates on the Raider? I haven't had time to really play with the G16 9955HX3D + 5070ti but it's still sitting on the table waiting 🤣

Crazy how when Nvidia needed the PC market and enthusiasts they were there with all the crazy ideas and letting AIBs do what they wanted because they needed the market share and us, yet as soon as they capped off dominating since Pascal, they have slowly but sure tightened and locked down all that made them great. Jensen really gets on stage and spews that tripe and expects people to eat it up. But alas, the golden rule continues to be in effect and right now the Nvidia product stack reigns surpreme. It very much reminds me of Intel at the top forcing new chipsets every cycle, rehashing old nodes and technologies, spoon feeding us performance increases and locking out features to maximize profits expecting brand loyalty till AMD came in guns blazing and now Intel has had no choice but to bring it all to the table each and every cycle and now they're going to offer socket longevity because, well, they have to to hope to recapture and hold consumers. Heck, we're even getting bLLC to compact X3D! This all reminds me when AMD legitimately scared Nvidia with the 6000 series and lost market share. With that fear lighting their fire, 4000 series brought everything they had to the table full tilt (and a cost to match) and they just left AMD in the dust. 5000 series has extended that overall performance lead, but again, all reports point to RDNA 5 being a wrecking ball...... Being fancy for the sake of being fancy........ problematic connector and the revised edition still problematic. Maybe third time is the charm?

Despite Nvidia trying their damndest to kill off SLI, dual GPUs are all the rage now for offloading rendering and especially in AI boxes with software designed to patchwork them together quite effectively. Another technology that will require more and more powerful cards ala early RT days and justify future purchases.....at higher costs of course at this rate. All reports really do lean towards AMD producing a monster on the level of the 6000 vs 3000 series days. Fingers crossed! Sucks but still a good set of sticks for someone looking to run 1:1 6000 tight. At 6000, tuned, do they run better than your other sticks tuned at 6000? That is where they could shine. I saw that price for the 2x32GB C26 kit and ran away as fast as I could! 🤣 Oh I know you will!! 😁 Hmmm, I have never ever heard of that place, but it seems to be a good used market front. That's a good price on those sticks and they should be a good bit better than your current sticks at 1:1 nice find!

Yeah that is out of my neck of the woods at 1.7 VDD. I stick to 1.5 max and let the chips fall where they may still #2 at 1:1 is a 6400 beast. If you're hell bent on 1:1 that is the way I would go. Curious to see what #1 and #2 do at 8000+ maxed out for some H2H "which one stays and which one goes," game show elimination. 🤣

OMG..... 🤣

Lunacy pricing.... 5060ti costs more than the 9070xt in the US atm retail pricing....just how much Nvidia doesn't care as consumers swirl the "edge computing" toilet bowl atm.... 62% faster at 1440p. 71% faster at 4k, costs less yet you know people will keep scarfing up the Nvidia brand. PC Partpicker: Cheapest 9070xt = $739.99 (~23% over MSRP $599.99 launch price) Cheapest 5060ti 16gb = $749.99 (~75%+ over MSRP $429.99 launch price) ------------------------------------------------------- Hmmmm..... Well 6400 vs 8000 you're not going to see much difference on average so you're right there. I talked about that over a year or so ago but 8000 is a cake and eat it too frequency to run. I kept going back and forth testing 6400 1:1 c30 and 8000 1:2 c36 and my final conclusion (YMMV of course) was that at worse, 6400 = 8000 and at best it beats on a varying scale so if your system supports it, especially running dual CCD which I don't even run on desktop, I am hard pressed to finalize around 6400. I would settle in and test at 8000 across the board and see which chip is fully stable at 8000c34 and weigh that equally, if not higher than 1:1 testing at this point, but again, that's just me. That is also one reason I've been jumping through so many hoops lately testing memory cooling the last 5-6 months because of the 270k adventures and finally adopting 8000+ on AM5. All setups going forward, after having stuck to 1:1 for over a year, will be 8000+ and those temps are getting juicy. On the other hand, if you're fully locked into wanting to run 2167/6400 tuned to the gills, CPU#2 would be the direction I would go too because it just seems made for 1:1. I'm hoping you get to eat your cake and eat it too and #2 also turns out to be the best at 8000. But man that not hitting 2200 is a (mild) kick in the teeth. The real suck out scenario is if CPU #2 is a beast at 1:1 and CPU #1 is a beast at 1:2. In that scenario, I'm picking #1 all day everyday. As for overclocking, I was pulling teeth and went with per CCD not per core like @Mr. Fox because it was quicker and you know if your weakest core is rock solid, the rest of them are too even with ambient changes. Last time I did per core was my 5800x and that took forever and even then I would sometimes get little instability and crashes. I was content with per CCD but of course you're not extracting the most performance possible so there's a trade off. Yeah, I don't know if I could let the 2200 chip go......

I am fully expecting you to snag one ASAP! 🤣 I did the exact opposite and we downsized from a monster sized house with a huge basement to a double wide since it is only the two of us now and the one thing I miss the most is the massive basement which was all mine and where I had multiple tables set up everywhere with room to spare. I had an "E" configuration with five 8' tables and closet space galore with a Vax 11/780 in the other corner I had picked up from the University of Penn they were tossing out back in the day. True fact, when we sold the house we left it in the basement because that thing was monstrous in size! It is the only reason I would entertain buying a small house with a same sized basement included is to set up my man cave again.

Loosen the timings and check 8000->8400 with TM5 Extreme and/or Karhu to see which one can truly stretch its wings the best unhindered by memory timing constraints but IMC or potentially MB constraints. See if 8200 can pass and if so maybe one of them can do an 8400 run. Once you see how high you can fly out the gate, then go back and start tightening up profiles.

First of all, congratz on getting two great chips! Second, have you tested their core strength and not just the IMC? Third, How are you testing? This is a situation I would throw Karhu at both of them and see what rises to the top for their IMCs Lastly, quite honestly to me, 8000+ is end game now for AM5 if only marginally so because while 6600 is nice the real end goal for AM5 is 2200+ and 8000+. I've upended all my timings and settings centered around 8000+ because at worst it performs the same as 6400 while at best it actually does a smidge better. The problem also is if you can only run 2167 then that throws out of sync a bit matching 2200:6600. Test both of them at 8000, 8200, 8400 and see which one rises to the top next. Real testing TM5 Extreme at least or Karhu if you want that end game verdict of where they stand realistically. I'm absolutely fine with powerhouse versions and this is exactly what I plan to pick up their 50 core + monstrosity. But with that being said, as long as they attach that cache to their 7 and 5 skus (or whatever they're going to call them), that will more than cover competing against AMD's 10850X3D (or whatever it ends up being called) along with the 9850X3D. I expect Intel to take the same approach as they did with Socket 1700 and to a degree 1851 and just come out blazing.... Keeping the same socket will also do much for longevity and good will. They finally realized the best way to keep customers locked in is making upgrades as easy as dropping in a new CPU. AMD figured this out a long time ago especially if your objective is to, well, sell CPUs?

I enjoyed this video. He states a lot of known truths from memory makers purposely constricting supply. Honestly I don't blame them with past precedent showing over production = collapsed markets. I think they know the AI market is tenuous at this point and to go all in could result in a massive collapse. I do think they are spring boarding off of the AI bubble with tightly controlled production to um "re-imagine" the consumer market pricing by overly constricting supply. There has to be a balance in there somewhere. Time will tell because right now memory is in stock everywhere in droves. Consumers aren't buying at these prices. Nvidia marginalizing gaming into edge computing because AI is the big thing right now I hope it bites them in the arse down the road but I suspect it won't because Nvidia is branching out into serious CPU production and other markets. Nvidia will be the final push that starts to seriously effect Intel and AMD's CPU production segment. All I'm saying is Intel better stay the course with their GPU division like AMD has to fall back upon. They need a two pronged attack while they continue to figure out a way to get their legacy x86 architecture modernized to compete against the M and M like chips coming into their own. "The craft builder is not really going anywhere because their reason to build was never the price and you can't price out a reason that was never about price." This is spot on. We build and tinker for the love of it. His comparison of craft builders vs necessity builders was very valid. He is also on point with the crowd that buys pre built because they just want to plug it in and go and how many who used to like to tinker have transitioned to this crowd. I have SO many life long friends and gaming buddies who used to get in the trenches to build and tweak their rigs to play Quake and Unreal Tournament and early days of WoW who just don't have the time or updated skill sets to build and tweak anymore and when problems arise they get immediately flustered and frustrated and usually turn to me for help where before we shared knowledge and ate our own chow. I think I have one friend left from the early days (circa 90s) who still insists on building his own rig. The rest either go pre-built, a laptop or have switched to Apple. ------------------------------------------------------------------------------------ Anyone who doesn't think 13th and 14th can still bring the pain where applicable is in denial. Dufus is stuck on bad testing and data gathering while asking us to "trust me bro" on the "feel" of games and a lack of consistency. I'm not saying there isn't some validity to what he is claiming, but he has built his entire brand around dry humping Socket 1700. But, as someone who still runs all three and routinely swaps and tests with them even now, Socket 1700 is showing its age. The most recent updates to Agesa (ugh) and gaming as a whole, all that dippity dip dip in WoW is basically gone at this point and remember, I was one of the early ones who was calling it out in the old forums and maybe even here with the 7800X3D vs 13900KS in WoW where there was some absolutely atrocious spots of it that are now gone. Even when I go back to my old "dip" spots to test that is streaming in assets at a high volume, the dip is no longer present. Best case, Socket 1700 is still a ripper in games that can't take advantage of X3D and doesn't pull a lot of power. Worse case, it is a very hot and heavy running CPU that gets left in the dust in games that love X3D while pulling 200w+. Intel badly needs a light and nimble 8 core / 16 thread large cache enabled CPU that doesn't suck down a metric ton of power to game and own. That's it. No more ; no less. I do think we will get that with Nova Lake as even Intel finally realizes they can't compete overall without a big 'ole chunk of cache. Nova Lake new architecture + bLLC is going to be an absolute wrecking ball. It is literally the #1 reason I am holding onto these Corsair Vengeance 9200 sticks to pick up a MB + CPU as soon as I get the ok from @Talon 🤣 ------------------------------------------------------------ Nvidia keeps pushing up pricing to protect margins and profit because quite honestly if the margins aren't there they won't stay engaged in even supplying to the gaming sector since we're now a blip in edge computing to them. Their objective is to keep pushing AI till they're blue in the face.... What's the alternative? People aren't buying 9070xt even though it is competitive against both the 5070ti and 5080. If the current pricing travesties won't push buyers to AMD, nothing will.

Temps are even better by 3c lower after a few adjustments down to ~71c vs 9800X3D sitting at ~80c. Not apples to apples though. #1. I didn't follow my own classic LM rule and actually apply a little extra on first application versus subsequent touch ups which I use less. I saw a few flat/dull spots and just added a touch more and reflowed the surface amount and redistributed on the heat spreader. #2. I added nylon flexible washers to the bottom plate. Something I've done on the regular in the past to not only compensate for any potential Z-height height issues (if any) but more importantly give a little flexibility to tamping down screws to auto compensate for any thermal expansion if I'm not using the ILM which I have sometimes removed for better/direct pairing and relied on hand cranked mounting pressure with a torque adjustable driver. I had initially opted to bypass this step but added it back in this time around. It provides more value on laptops that support desktop CPUs overall with all the various pairing issues and such. I ordered a few custom cut shims ~2mm that will arrive next week for some comparison data vs no washers vs several AIOs especially the ID Cooling Frost Flow x 280 which has zero compensation/extended range when tamping down the CPU block vs my EK CR240 or Lian Li II Performance Lite 360 which both give extra headroom for pairing and can somewhat compensate for the loss of the 2mm z-height with the TG Heatspreader. I do think TG should have offered a normal height heatspreader or at least present their data showing core temp differentials between both versions tested in house. Maybe it is floating around out there on the Internetz somewhere. -------------------------------------------------------------------------- What a great advertisement for the TG WV II Pro....... seriously, Roman sits around and thinks of the stuff HE would want as an enthusiast, former high level OCer and just lover of tech. Not sure why he doesn't have or used GPU Tweak III to see how it responded.... Watching it immediately pick up that load imbalance and spit out a warning was nice to see. Hook in the auto off wire function and it would even immediately shut down your system for you too. That's two be quiet! PSUs involved in meltdowns in the matter of a few days. One using the 12vhpwr and the other using 4x PCIe 8 pin. I don't think it is be quiet's fault at all. Just an odd coincidence but more importantly burns and failures completely away from the 12vhpwr connector on the GPU itself but instead the PSU and/or PSU cabling. I'm sure they're going into damage control mode though as a precaution. be quiet Dark Power 13 (This is HUB, I'll assume the 1600w model) be quiet! Dark Power 13 (Club386 burned at both ends) be quiet Straight Power 11 Platinum 1200w (Tech Notice) (time stamped at PSU ID) Just a trash connector....it really is......

So glad everything is working out this time around! Last time, you had three points of failure to try and diagnose (MB, Mem, CPU), and returning it all was the right call and paid dividends nicely. 6600 is super golden as is 8800 like on Sugi's setup which is bonkers. Makes ya wonder about the second one you have at this point..... ------------------------------------------------------------------------------------ This 9850X3D can do 6600 but fails during TM5. 2233 flat out won't boot but 2200 on stock is just fine. Temps are the problem as they were fine the first few days (~74c under CB23 looping load) and now are hitting 95+ from the jump so I will need to pop the Thermal Grizzly Heatspreader and check the liquid metal application. I needed to install some nylon washers on the backplate anyhow to help with the ~2mm z-height differential anyhow for more AIO options as the eK cr240 Nucleus was the only model that allows a deeper hand torque to somewhat compensate but once offset temps will plummet even moreso. ----------------------------------------------------------------------------------- Still fighting the urge to install my 270k in my main rig and eat my own chow till Nova or Zen 6 since even on this Dark Hero with a 9850X3D, it is still more of the same just a bit better..... 🤣 I know if installed as my main, it will get real time tuning and TLC....

Looks awesome @Reciever! You've really switched into that next gear with hardware and cooling! Do you have another platform to test it with the block on real quickly to see if it boots with a display out?

That makes sense and good to know they should work on the MSI and Gogabyte boards. The KingBank sticks being problematic on your former X870E Apex adds a nice bit of validation to the problem since we ordered them within a week or two of each other from da zon.... That also means that even with a BIOS update the HP 35L most likely would work with them as HP uses their own custom boards and cancerware. My problem is/was having 2x Asrock boards and 2x Asus boards. Asrock having problems with their sticks from the jump and Asus developing issues after this latest Agesa update (could be more even on the slightly older BIOS, but I didn't FAFO enough to see)....yet Corsair, Teamgroup and Kingston (HP) Rage sticks keep on truckin'..... I could get one stick of the KingBank to kinda, sorta sporadically post on the Dark Hero but then it would immediately freeze at the Asus logo screen. On the other hand, KingBank absolutely flawless on Intel and an absolute joy to use. Was able to get my 14900KS to post 8800 and the 270k pass TM5 up to 8933 in G4.... -------------------------------------------------------- Nice to see you in the 18k club! More importantly not needing the external hardware but software (as it should be) to use your GPU properly. If it turns out Nvidia is purposely going out of their way to neuter mVolt+, we'll be stuck with older drivers going forward just to be able to use our GPUs as we want.

So the AM5 and KingBank issues persist but I'm pretty sure I've narrowed down what is going on even further. I tried to pop them in the Dark Hero for heatsink / air cooling testing only to be met almost immediately with an "A6" error code and a bright yellow dram error light. AIO fans kept revving up and down in the oddest pattern. I know they work in my x870e Hero since I just posted and booted with them last month right? So I popped them in there and exact "A6" error code and bright yellow dram error light along with AIO fans revving up and down. I drag out the Microcenter combo deal Asrock B850 board and of course they don't post in there as expected but why not throw another data point on the pile..... They did post in the HP Omen 35L w/ 8400f but that drive is wiped atm so no further testing..... As a final checksum, I pop them into the 270k build and of course they not only boot right up but blew through TM5 looking at me like I'm the nut job in this scenario 🤣 I KNOW these sticks posted and booted in my x870e Hero last month, but in the interim I had updated my BIOS as I was running 2103 from March. Updating to 2402 updated Agesa from 1.3.0.0a to 1.3.0.1b. The HP is actually running a BIOS from earlier this year. The Dark Hero I updated upon initial install as it was running 0502 to the newest so it too is running 1.3.0.1b. I'm thinking if/when HP updates to a newer BIOS supporting 1.3.0.1b those sticks might stop working in that system too. On a whim, I rolled back the Asrock BIOS from 4.43 to 4.10 which supports 1.3.0.0a and the sticks still gave me a big middle finger but with the reported incompatibilities with these sticks on Asrock boards, I wasn't expecting them to work. Last gasp I popped in my TG sticks and set a loose 6000, 8000 and JEDEC profile and saved all three and sat there loading in each profile on the TG sticks then trying to get the KingBank sticks to boot using those known, working profiles but nope right to "A6" DRAM error light on both boards..... Can't roll back on the Asus boards but at this point I wouldn't anyhow. I'm just done with these KingBank sticks on AM5. I'll stick with the Vengeance 9200 and TG 8200 sticks which continue to work like butter no matter what. I picked them up June 2025 and outside of some quick testing on the x870e Hero at the time, they were primarily used in benching and testing my SP109 14900KS before I sold off the Z790i lightning and moved the 14900KS into the wife's system. It is only recently they're getting attention as they were slated for my SFF build. I know in my 270k they could post up to 8933 (maybe squeeze more out of them better cooled) so I'm just going to tune them up there around 8400-8600 and keep it moving or look to trade them for a set of 6000-6400 CL26 sticks or similar. Fun times!