Talon

Intel WILL improve their GPU performance more and faster over time versus the established big players. Not because they are better, but because they have so much untapped potential to still squeeze versus the others that are milking all they can out of the old cow. I really want an A770 just to tinker. I've always liked Tom, such a force behind many of Nvidia's achievements. I'm glad he's helping drive the ship over at Intel.

Apparently migrating the servers, supposed to be done in the next couple of days.

I tested 3. I got SP99, SP101 and SP106. I gave my brother my SP99. Kept the SP101 and SP106. I'm not bothering with further binning and doubt I'll go KS. I am very happy with any 13900K and memory OC is the META now. Core clocks bring minimal gains for my use case and default is just fine. Memory OC brings much bigger gains. I grabbed a Z790 Apex and 8000 CL36 is now easy. Hoping EVGA gets their shit together and gets a decent BIOS out soon, looking like 2.04 is trash. If they do fix it, I'll rebuild the second rig with the Dark board Right now I'm sitting on some hardware, but I'm finally selling stuff off again. Thinking about dumping my laptop too since it's not been used in months at this point.

What happened to your recent golden chip you just got from BB? Did you sell that one too?

YEP I've brought this error up with EVGA and Asus, I've had it happen on both brands. It's an A-Die bug or something. You push too hard and then it just won't post anymore. Then nothing works to fix it. The only thing that has worked for me is to clear CMOS and then put in M-Die sticks. Absolutely insane that it could lead to a soft brick. Had I not had the other sticks I would have been dead in the water or possibly switched BIOS. But most casual users would have been lost. I saw someone else post with this exact bug/issue.

I'm sure you're on the latest BIOS, but still EVGA has yet to tune the BIOS for proper A-Die support IMO. I think they're either saving that for Z790 release or using it as a way to get people to buy the Z790 Dark and that annoys me. Asus is releasing BIOS updates almost daily at this point, constantly tuning for A-Die, to the point that I have 7200 CL34 A-Die stable on my Z690 Hero 4 dimm board plug/play XMP. That is just sad that such a high end 2 dimm board is struggling so badly. Try 1.26 VDDQ IVR TX/CPU VDD/ CPU VDDQ TX. I can't remember what it's called in the EVGA BIOS right now. After I did this I was able to boot well into 8000+ but never stable above 7800 CL36.

They work great for quick CPU test, for example just making sure the system is up and running, checking an SP score, etc. No mess, not wasting tons of cash on expensive paste. You could easily use them 24/7 if not running high clocks/voltage and benching. Like if you were just to build a gaming system, you 100% use the thermal pad and not have overheating issues. But for best temps and highest perf, of course good paste is the way to go.

100% agree. After using this cable on a couple 40 series cards, I've had zero issues with the cable other than being semi difficult to plug in and unplug. I mentioned that a couple weeks ago when I got my cablemod cable I noticed how difficult it could be to just seat the cable fully. And look and what came out, it actually is people failing to ensure a properly seated cable. In my opinion, people building computers and no less responsibility to ensure it's done properly than say someone installing a new lighting fixture. If you manage to do that incorrectly you could just as easily burn down your home. Should those people be allowed to turn around and sue the company that sold the lighting fixture? But Nvidia and the cable spec should have made this easier IMO.

Tried and tried to get my Z690 Dark to get stable with A-Die at really high clocks. Even though they released a BIOS that finally got it booting high memory frequency, above 7800 CL36 I could not stabilize. Maybe they'll drop a BIOS that fixes that, maybe they won't and will "encourage" customers to buy the 790 to get stable high A-Die clocks. The Apex stabilized at 8000 CL36 at stock 1.4v for DRAM and autos on everything else with XMP Tweaked Preset. Played an hour on Warzone II and 2042 and no issues after testing ram. I've thrown a bit more voltage at the 8200 CL36 I'm testing now, no idea since I just got it all setup this evening.

Still a banger CPU. Especially with 4000 Gear 1. Not sure if you picked up a CPU yet, but I think 13900KF at Amazon was pretty cheap the other day. https://www.amazon.com/Intel-i9-13900KF-Desktop-Processor-P-cores/dp/B0BCFM3CJ4/ref=sr_1_1?crid=3UZALL1N1YHSU&keywords=13900kf&qid=1668630132&sprefix=13900kf%2Caps%2C116&sr=8-1&ufe=app_do%3Aamzn1.fos.4dd97f68-284f-40f5-a6f1-1e5b3de13370

I love the beginning of the video where the cable is basically hanging out in order to get it to burn up. Failure Rate of 0.05 to 0.1%. And then of that small percentage, what percentage is because some people drew the short stick in the thought department.

Absolutely, good stuff for non RTX owners. I tested both yesterday when it came out and the DLSS image quality is better and more stable. The FSR 2.1 flickers more often and it's especially bad when driving at times, heavy ghosting/trailing when the camera jerks or has a chase camera angle. This is especially noticeable you see in person vs YouTube heavy compression. Still if I had an older GPU, with no RTX, I would choose FSR 2.1 over native.

Damn do most cards stop at +1000? My Strix runs out of room at +1900 with +1850 being the limit. But honestly the chip and memory chips are all a lottery, no brand is better than the other this gen. Card to card variance is real and you're paying for looks, cooling, and warranty reliability.

@tps3443 This is what you were supposed to ask for in those Freebies lol. 😁

Yikes. I am so far so good, and I always shut down my rig when not in use. I still believe these people are not plugging the cable in all the way, it is very easy to do. When I plugged in my CableMod cable I realized/saw just how easy it is to think you have fully seated the cable when in fact the cable is not seated at all and easily pulls right out. This has been shown by trusted tech outlets are a possible and likely cause. It shoudln't be this easy to mistake and should be addressed in some way or public statement. Really happy I have the Microcenter replacement warranty right now. No need to wait for RMA or anything like that. Walk in with issue, walk out with new card or $$$$. Maybe mine will fail when the 4090 Ti drops for similar price haha.

I love watching this guys videos. Entertaining and informative stuff.

AMD confirms 7900 XTX is 4080 competitor, not 4090. Damn crazy to hear AMD come right out and say they have no answer for the 4090. If it’s competitive the 4080 that will be good for all, and I see Nvidia dropping prices a tad in the near future, at least on the mid range cards since AMD has bowed out of the high end now.

Interesting, maybe I need to go higher then myself.

Nice and thats on a Z690 Hero. I doubt you need that high of VCCSA. Honestly think 1.20-1.25 should be more than enough. Maybe try and lower that?

@tps3443 @johnksss @Mr. Fox Have you tried lowering your CPU VDDQ/IVR TX Voltage/VDDQ TX Voltage. Honestly we need a same naming scheme across brands. For me to get higher A-Die stable clocks/push it hard you need to actually heavily decrease this voltage which is opposite of what you would think. For me 1.18-1.26v is the sweet spot, with mine liking 1.26v for 7800 CL36. With M-Die it was opposite and higher VDDQ TX helped to a point, usually around 1.375v worked for higher than 7000. I think I read somewhere that you should run it 10% below the VDD/VDDQ on the DRAM. So say you're running 1400mV-1435mV for A-Die, that would put you in the ballpark of 1260mV for VDDQ TX. Likewise if you're running higher voltage for M-Die, say 1500-1550mV, again it puts you around 1350-1395mV.

Again, those cables they are packaging are not native 12VHPWR cables, they are cables that are adapted to meet the spec using existing power supplies, so potentially another point of failure if we use the same analogy you used for a CableMod cable. You do realize the CableMod cable is doing exactly the same thing the Corsiar or Seasonic adapted cables are doing, right? The difference is that CableMod decided to go for more redundancy and lighten the load per 8pin cable vs forcing 300w down the cable that is actually SPEC'd for just 150w. https://wccftech.com/atx-3-0-12vhpwr-gen-5-connector-major-safety-risk-using-adapter-confirms-pci-sig/ Obviously we all know the cables can handle far more, but why push it when you can not be cheap and provide a quality cable? Also they know some XOC community will want to push even beyond 600w, and future cards could easily draw beyond that. So again, redundancy and headroom is the way to go. The only PSUs that have native 12VHPWR are ATX 3.0 PSUs which have a 12VHPWR connection on the PSU, they do not use 8pin to the 16 pin.

Congrats!!! I just celebrated 9 years with the wife last month.

The farthest I could push my M-Die (so far) 7800 CL36 with auto timings. Unfortunately not stable. 7200 CL34 just passed 30 minutes stability testing no issues at stock 1.435v. Insane value dimms.

I'm on a DDR5 testing kick lately. Currently testing my TeamGroup 5600 CL46 $90.99 (Amazon) SK Hynix-M at 7200 CL34 1.435 VDD/VDDQ. They booted and ran AIDA at 7600 CL36 but cannot stabilize. Right now running a stab test at 7200 CL34 and looking good. For $182 32gb of these sticks are an phenomenal value play. A-Die is cool and all, but honestly this is going to provide you with 99% of the perf on Z690. Z790 might be better for A-Die. https://www.amazon.com/dp/B0BC8NHWJZ?psc=1&ref=ppx_yo2ov_dt_b_product_details Edit: Hell even 7800 CL36 is booting on these cheap-0 sticks lol. Runs AIDA64. Loosened the TRFC a bit and set voltage to 1.5v.

Ok so it seems we might be at the limit of Z690 then.