re: PC Discussion - Gaming, Performance and Enthusiasts

HD7970 got nothing on the GOAT 6600GT AGP

Yes, especially new architectures. Same thing with first gen Ryzen and first gen RDNA.

Hell even new phones, I had no desire to get the first Gen Fold knowing it was gonna have issues. Waited for 2nd Gen with huge improvements and glad I did.

Same with new cars too, avoid first year of a new model or redesign.

quote:

---

JayzTwoCents is saying it is how the AIB cards are designed that is the issue. POSCAPs are the issue instead of using MLCC's, AIB's cheaping out under spec.

At the end he says he doesn't expect AIB's to retool and remanufacture the cards and he says no. Says they'll release a VBIOS update with lower clocks instead. Also: avoid ZOTAC at all costs.

---

Thats not an Nvidia problem. Thats a major frickup by the AIBs if true.

Also - why avoid Zotac? They've had pretty good build quality cards in the past; I've had no issues with them.

VBIOS update is the most cost effective measure. I'd imagine the AIBs would release a Rev 1.1/2.0 version of the impacted cards at some point if this blows up further.

quote:

---

Asus and EVGA only ones confirmed so far to be properly built.

EVGA has just confirmed they discovered this issue in QC and fixed it. This is a super bad look for AIB's.

---

I'm guessing the AIBs were forced to reduced costs due to product tiers and marketing, and also management. Not that uncommon, but still a bit shocking. The engineers and technicians of these companies are really smart, but they are overruled in many situations.

Based on reviews, the ASUS TUF Gaming, and the EVGA XC3 series seem to be the best cards at the moment.

Wonder any AIB will release a blower-style, or an ITX form factor card.

quote:

---

If it’s still hitting the advertised boost clocks that the AIB claims, then in that sense it’s not “bad”.

---

Exact opposite, the boost clocks are what is causing the crashes. They can't handle the boost due to the poor cheap POSCAPS. That is exactly what is causing the issue, and why he says he doesn't think AIB's will fix the cards but instead issue a VBIOS update lowering the boost clocks.


HardwareUnboxed is completely refuting this claim by Jayz and his testing however. They say all the cards crash the same amount regardless of the caps used and that the real issue is still unknown. LINK

quote:

---

Hardware Unboxed @HardwareUnboxed 4h

The crashing with the RTX 3080 cards doesn’t appear to be down to the caps used, which is why we haven’t made a video yet, we don’t know the issue. What we do know is the FE and TUF Gaming models crash just as much as other models and they use MLCC’s.

---

quote:

---

Exact opposite, the boost clocks are what is causing the crashes. They can't handle the boost due to the poor cheap POSCAPS. That is exactly what is causing the issue, and why he says he doesn't think AIB's will fix the cards but instead issue a VBIOS update lowering the boost clocks.

---

Hmm, maybe I’m misunderstanding the issue but in the video Jayz explained that the cards were still hitting advertised/guaranteed clocks, but that the issue was with GPU boost thinking it had headroom and then the cards crashing because of the tables in the VBIOS that dictate GPU boost based on power/temp inputs

Oh my if true yeah I'm upgrading.

quote:

---

AMD Ryzen 9 5900X “Vermeer” 12 Core & 24 Thread Zen 3 CPU Allegedly Up To 5 GHz With 150W TDP


As for the alleged specifications, the AMD Ryzen 9 5900X is said to feature an IPC improvement of up to 20%. One of the main areas of improvement where AMD has put lots of effort into tuning on Zen 3 is the clock frequencies. As such, the AMD Ryzen 9 5900X is said to reach boost clocks of up to 5 GHz.

---

Really though it's only 25W more than 10900K with 2 more cores and 4 more threads.

quote:

---

Anything on 5700x/5800x? Hopefully one of them is a 10 core.

---

Extremely doubtful they go 10 core, would mean a major redesign. It will all be the same cores/threads just boosted clocks and better IPC - although I tend to doubt the 20% claim, that's a bit ridiculous on a mostly refresh. Going from zen to zen+ and that was 14nm to 12nm, not 7nm to 7nm like zen 2 to zen 3, was way less than 20% IPC improvement, think iit was what, 5% or less? I think most of the gains will simply come to clock speed increase, especially if they can get best models up to 5ghz out of gate.

Probably something like this just by guessing:

3600xt (6c/12t) 4.5ghz boost -> 5600x (6c/12t) 4.7ghz boost

3800xt (8c/16t) 4.7ghz boost -> 5800x (8c/16t) 4.9ghz boost

3900xt (12c/24t) 4.7ghz boost -> 5900x (12c/24t) 5.0ghz boost

3950x (16c/32t) 4.7ghz boost -> 5950x (16c/32t) 5.0ghz boost

Increase base clocks by about 200mhz across range too.