*Official Benchmark Thread* - Post it here or it didn't happen :D

[electrosoft]

4 hours ago, Rage Set said:

 

His shots at Steve of GamersNexus and Steve of HUB are funny and ironic. FC claims they took their issues with these companies personally, but it almost seems he has a personally problem with AMD. That commenter hit the nail on the head.

 

 

I thought the same thing. 🙂 That comment was spot on.

 

3 hours ago, Mr. Fox said:

Yeah, the memory speed used by the GPU matters a ton. A whole lot more than system memory. 2667 versus 3200 would definitely be measurable. This is one of the reasons the flagship Radeon cards can't compete with flagship GeForce cards. AMD still use GDDR6 versus the far superior GDDR6X video memory. In some conditions the performance difference is tremendous. The slower speed and decreased bandwidth of GDDR6 can often present a remarkable performance bottleneck.

 

 

I imagine AMD had no idea Nvidia's halo product was going to be so powerful. If Nvidia topped out with the 4080, it would be 3000 vs 6000 all over again (7900xtx and 4080 trade blows, AMD loses (again) in RT). 4090 just leveled everything.

 

39 minutes ago, tps3443 said:

Anyone try one of these liquid devil 7900XTX? They cost $1,399.99. I think it’s cool looking. 
 

https://www.microcenter.com/product/664356/powercolor-amd-radeon-rx-7900-xtx-liquid-devil-overclocked-liquid-cooled-24gb-gddr6-pcie-40-graphics-card

 

How many in this group even have a 7900xtx? Me (again) and @Raiderman? They also have an air cooled model and they had a limited edition run too. I always liked the look of their cards.

 

De8auer took a look at it a few months ago. I like the look and aesthetics.

 

 

[ryan]

@Mr. Fox in about 2 weeks ill send you some money for the score that beat my uber Xe. I'm still in shock, can't believe you beat it, and only by like 400 000 points. I dont have paypal but ill set it up. I'm thinking of starting a draw, I get free money so why not give it away for free lol. ah good times. loved your benchmarks, some of the highest scores I have ever seen in the respective benchmark. when I get my Hp omen I plan on flashing the vbios, I hope its as easy as it was back in the 6970m days.

 

also as @electrosoftmentioned the 4090 mops the floors. clean win even when the 7900xtx is overclocked, amd is going to need to produce a killer 8900xtx or something, not sure how they stay in business with everyone buying nvidia, it's got to cost alot designing and engineering these cards

 

[Mr. Fox]

8 hours ago, ryan said:

@Mr. Fox in about 2 weeks ill send you some money for the score that beat my uber Xe. I'm still in shock, can't believe you beat it, and only by like 400 000 points. I dont have paypal but ill set it up. I'm thinking of starting a draw, I get free money so why not give it away for free lol. ah good times. loved your benchmarks, some of the highest scores I have ever seen in the respective benchmark. when I get my Hp omen I plan on flashing the vbios, I hope its as easy as it was back in the 6970m days.

 

also as @electrosoftmentioned the 4090 mops the floors. clean win even when the 7900xtx is overclocked, amd is going to need to produce a killer 8900xtx or something, not sure how they stay in business with everyone buying nvidia, it's got to cost alot designing and engineering these cards

I am glad you liked the benchmark scores. I had fun for the first time in a while last night. No need to send money. I don't know if you are being silly or not, but you have demonstrated generosity more than once. That is a great character trait. If you buy the NUC I'll count that as payment for the benchmark scores, LOL. 🤣

 

Part of it is Intel and NVIDIA are very sore losers and exhaust financial resources on winning. That includes the people they hire, and what seems to be an indication of "winning is more important than how you play the game" mentality. The other part of it is AMD seems to lack experience and drive. They talk like they want to win, but are not willing to do what it takes to score. Using GDDR6 instead of GDDR6X is an example of that. Just that one change would go a long way in closing the performance gap between 7900 XTX and 4090. Using GDDR6 on their flagship GPU in 2023 is just lame.

 

They've been playing in the console sandbox for too long. They're out of touch with PC gaming and overclocking enthusiasts. They hire people that aren't advocates for overclocking enthusiasts and don't really have a clue what is important to us. And, they may not care. Their drivers leave something to be desired.

[cylix]

@electrosoft Looks like the Red Devil has some problems with the heat

https://videocardz.com/newz/powercolor-rx-7900-xtx-red-devil-cards-may-suffer-from-high-gpu-temperatures-due-to-incorrect-application-of-thermal-compound

 

@ryan 7900xtx can beat 4090 in 3dmark but you need some serious OV, 3,33 GHz and 700 W power consumation 😆

 

AMD needs to work on the power consumation of the cards. Potential is there just need to work the quirks out

https://www.tomshardware.com/news/amd-rx-7900-xtx-matches-rtx-4090-at-700w

[Mr. Fox]

2 hours ago, cylix said:

@ryan 7900xtx can beat 4090 in 3dmark but you need some serious OV, 3,33 GHz and 700 W power consumation 😆

 

AMD needs to work on the power consumation of the cards. Potential is there just need to work the quirks out

https://www.tomshardware.com/news/amd-rx-7900-xtx-matches-rtx-4090-at-700w

While it sounds cynical on the surface, the fact of the matter is that has been the excuse/reason for too many years. Lots of near misses and always the same or similar explanation for it. Too hot, too much power/voltage needed, just need to work on better drivers, just need to work on better firmware, just need better support from partnering hardware vendors or game developers. I think they would if they knew how. I think the real problem is they don't know how to win, and don't hire people that know how to win. Part of the time is was lack of financial resources, but that's no longer an issue... so it leads me back to not knowing how, and not really understanding what some of their prospective customers would require as prerequistes for entertaining the idea of changing brands.

[electrosoft]

1 hour ago, ryan said:

also as @electrosoftmentioned the 4090 mops the floors. clean win even when the 7900xtx is overclocked, amd is going to need to produce a killer 8900xtx or something, not sure how they stay in business with everyone buying nvidia, it's got to cost alot designing and engineering these cards

 

 

The 4090 is a monster but the 7900xtx in pockets has its moments. The real argument is cost per frame ($1600 vs $1000). If you're on a budget $600 is a major chunk of change but flat out best? 4090. Unless AMD has something up their sleeve with a 7950xtx that is more than the 6950xt was vs the 6900xt, Nvidia has to do nothing but just sit back and cruise...maybe introduce a 4080ti if needed to fill that gap and/or a 4090ti just for the icing on the cake.

 

It's been almost 5 months since my XFX 7900xtx adventures and I've had this Suprim Liquid X 4090 for over 4 months now and it is just a monster (2820-2835 boost out of the box, 3120 rock solid gaming/benching on stock BIOS, +1600 mem)  so no complaints. 7800X3D is still playing catch up but doing much better than my 12900k did.

 

I figure I'll give my favorite pick this generation of 7900xtx's another try and see.

 

 

4 minutes ago, cylix said:

@electrosoft Looks like the Red Devil has some problems with the heat

https://videocardz.com/newz/powercolor-rx-7900-xtx-red-devil-cards-may-suffer-from-high-gpu-temperatures-due-to-incorrect-application-of-thermal-compound

 

@ryan 7900xtx can beat 4090 in 3dmark but you need some serious OV, 3,33 GHz and 700 W power consumation 😆

 

AMD needs to work on the power consumation of the cards. Potential is there just need to work the quirks out

https://www.tomshardware.com/news/amd-rx-7900-xtx-matches-rtx-4090-at-700w

 

Yikes, such thermal application issues can make or break gpus, cpus right off the line. Well the good news is @tps3443 would crack that open ASAP and repaste it anyhow.

 

Shunt modding a 7900xtx in response to AMDs lock down and pulling 700w? I love it and it shows potential. 🙂 As alluded to above, I hope AMD brings a 7950xtx (7970xtx?) to market with a bit more punch to compete with the 4090. I've always had a soft spot for AMD and used their (well ATi initially) GPUs more often than not over the last 23 years since I purchased my first rage card and OC'd the snot out of it to play Deux Ex back in the day. Ran that 5700xt for over a year and a 7970 for 1.5yrs.

 

 

 

 

 

 

 

 

[ryan]

I wasnt joking. I don't mind. I'm a pretty nice guy. I live by the saying treat people how you want to be treated with 1 or 2 exceptions. I love it when you guys post your scores, I always backcheck to see where they are in the big scheme of things. I would also like to donate to NBT but they have not setup a payment method yet even though it was requested a year ago. If your nucs available ill be happy but dont hold back because of me if you need money for it u need money for it.

 

 

managing 40fps at lowest 1080p with thps 1 + 2

 

heres maximum setting no aa low shadows. it seems to get 40 no matter what I do graphics wise. cpu bottleneck?

 

 

its really impressive how far igpus have come, but I guess its more impressive how every year the dedicated chips go up 50 percent.

 

1080p is the new 720p. when igpus have the power for 1080p you know with a ded card you should be gunning for 4k

[electrosoft]

On 5/16/2023 at 12:15 PM, electrosoft said:

 

@Raiderman your hub replacement seemed to ship pretty quick, nice!

 

 

Still waiting on Cablemod to ship me out a new cable for my 4090 for the defective one they sold me. It's been almost two weeks now since the CSR, "put in the request," for a new cable to be shipped after having me physically cut my old cable in half and send a picture (great way to trust your customers).

 

Cablemod finally shipped my replacement out. It is coming directly from China.

 

 

Now I just have to set up sending in my AC LF 420 to Arctic for a new unit since mine is in the recall window. During the transition to AMD, swapping over to EVGA freed the monster 420. 🙂

 

 

[Papusan]

9 hours ago, electrosoft said:

As alluded to above, I hope AMD brings a 7950xtx (7970xtx?) to market with a bit more punch to compete with the 4090.

3x8 pins ain't enough. AMD will have to change the power connector. Even their best binned dies won't be enough to stay with old fashion 8pin connectors. + we know AMD cut corners where they can (GDDR6vram etc) then they have the cheapo man F. Azor that look at maximum profits over awesome. 

 

They can close the gap to 4090 but will never come on par this gen cards. Nvidia don't even need 4090Ti to stay on top as long AMD only look on the bottom line. 

 

10 hours ago, Mr. Fox said:

I think the real problem is they don't know how to win, and don't hire people that know how to win. Part of the time is was lack of financial resources, but that's no longer an issue...

 

Or change from... I think the real problem is they don't know how to win to....... I think the real problem is they don't want to win because that will increase costs. 

[electrosoft]

27 minutes ago, Papusan said:

3x8 pins ain't enough. AMD will have to change the power connector. Even their best binned dies won't be enough to stay with old fashion 8pin connectors. + we know AMD cut corners where they can (GDDR6vram etc) then they have the cheapo man F. Azor that look at maximum profits over awesome. 

 

They can close the gap to 4090 but will never come on par this gen cards. Nvidia don't even need 4090Ti to stay on top as long AMD only look on the bottom line. 

 

 

Or change from... I think the real problem is they don't know how to win to....... I think the real problem is they don't want to win because that will increase costs. 

 

I'm not advocating they try to offer a 700w card but positing the possibility of a card with an expanded chiplet design up their sleeve. 🙂

 

In reality, the 4090 is the only card that truly separates itself from the pack (and then some) as the 4080 and 7900xtx are pretty close in performance sans RT. It is last gen all over again except for that pesky 4090. 🤣

[Mr. Fox]

11 hours ago, Papusan said:

Or change from... I think the real problem is they don't know how to win to....... I think the real problem is they don't want to win because that will increase costs. 

It could be either one, but the outcome is the same either way. Not winning is not winning. Like RIcky Bobby's daddy always said...

11 hours ago, Papusan said:

3x8 pins ain't enough. AMD will have to change the power connector. Even their best binned dies won't be enough to stay with old fashion 8pin connectors. + we know AMD cut corners where they can (GDDR6vram etc) then they have the cheapo man F. Azor that look at maximum profits over awesome. 

Well, 3x8-pin was enough for the 3090 KPE pulling over 750W through 3x8-pin. I do not believe the 12VHPWR connector was necessary and I do not consider it to be a real improvement. The one advantage that I see is that it is one cable versus three.  But, for the same reason it is more fragile and prone to failure than the tried and true 8-pin cable. Pushing the same amount of power through a smaller connection and fewer wires, in a concentrated spot on the PCB, also makes  is hotter.

10 hours ago, electrosoft said:

I'm not advocating they try to offer a 700w card but positing the possibility of a card with an expanded chiplet design up their sleeve. 🙂

 

In reality, the 4090 is the only card that truly separates itself from the pack (and then some) as the 4080 and 7900xtx are pretty close in performance sans RT. It is last gen all over again except for that pesky 4090. 🤣

Nothing wrong with 700W. More power generally means more performance. When it consumes gobs of power, gets hot, but still doesn't deliver a commensurately high level of performance is where that becomes problematic.

22 hours ago, electrosoft said:

I imagine AMD had no idea Nvidia's halo product was going to be so powerful.

Having no idea about some things seems to be a trend for them.

[Papusan]

1 hour ago, Mr. Fox said:

Well, 3x8-pin was enough for the 3090 KPE pulling over 750W through 3x8-pin. I do not believe the 12VHPWR connector was necessary and I do not consider it to be a real improvement. The one advantage that I see is that it is one cable versus three.  But, for the same reason it is more fragile and prone to failure than the tried and true 8-pin cable. Pushing the same amount of power through a smaller connection and fewer wires, in a concentrated spot on ths PCB, also makes  is hotter.

Yep, no problem go outside the 8-pin PCI Express power connectors current rating specs if the GPU manufacturer will allov it. But this won't happen. Needed power is only possible in combination with the fragile 12VHPWR connector or 4x 8-pin PCI Express power connectors. And this extra adapter in the GPU package will increase the costs aka reduce AMDs profits. Also adding the 4th needed 8-pin PCI Express power connector on the PCB will increase costs. Going the cheapo route will always come at a cost.

 

Leadership and Efficiency - Perfected by AMD...............😎

 

1 hour ago, electrosoft said:

I'm not advocating they try to offer a 700w card but positing the possibility of a card with an expanded chiplet design up their sleeve. 🙂

 

A major change for the GPU package only to try coming 2nd will cost AMD a hell lot money. And there was for a reason they only added the XT over X for their older Ryzen Cpus. 

 

The AMD way... Being 2nd and keep up max profits.

 

Why is AMD not using GDDR6X? - PC Guide

[Mr. Fox]

8 hours ago, Papusan said:

Why is AMD not using GDDR6X? - PC Guide

As I mentioned before, there is a huge disconnect in their understanding and thought process. They do not understand their target market and they have confused console jockeys with PC enthusiasts. The alternative is they don't care or they are targeting a compromised group of consumers rather than PC enthusiasts.

 

Believing that GDDR6X doesn't offer a meaningful performance gain and that 90°C+ hotspot temperatures is OK paints a picture of a truly clueless and misinformed clown posse. They will remain in the shadow of NVIDIA forever unless these examples of poor judgment are mitigated.

[Talon]

 

Asus BIOS engineer showing Steve how to actually measure voltage. 

 

[Papusan]

4 hours ago, Mr. Fox said:

As I mentioned before, there is a huge disconnect in their understanding and thought process. They do not understand their target market and they have confused console jockeys with PC enthusiasts.

Also Nividia isn't able to understand own stupidity..... They only understand greed. Yep, no price hike but they castrated their 4060Ti cards with less Cuda cores and lower MEM Bus vs 3060Ti to cut costs. Easy make matching prices if you destroy own brand new graphics cards. 

 

And Nvidia want $100 USD for 8GB extra vram. How much does it cost Nvidia to buy it? Not much more but hey will use the more expensive Clam shell mode for 4060Ti. Thats retarded.

 

4060Ti. What a fantastic generation performance uplift. 15 percent faster performance than the 3060 Ti in games without the DLSS Frame Generation feature enabled. Up to 52% higher boost clocks but only 15 percent faster real performance... Yep, reducing CUDA cores and MEM Bus vs its predessor come at a cost. Yep, what an improvement. Rather a Joke. 

Nvidia introduces $399 RTX 4060 Ti and $299 4060 without introducing a price hike arstechnica.com

[ryan]

Epic games store actually has a pretty kickin game for free right now just caught it in the nick of time. death stranding is free and a great game to benchmark as it's well optimised and the performance to visuals trade off is one of the best.

 

anyone here news on GDDR7?

 

also @Papusan I was thrown back after watching the 3060 vs 4060 laptop comparison by jarrods tech, like a 10 percent bump for 300 bucks. its very frustrating seeing technology de-evolve or go into stagnation due to greed. we need more competition as it stands amd/intel/nvdia isn't enough

[Papusan]

33 minutes ago, ryan said:

also @Papusan I was thrown back after watching the 3060 vs 4060 laptop comparison by jarrods tech, like a 10 percent bump for 300 bucks. its very frustrating seeing technology de-evolve or go into stagnation due to greed. we need more competition as it stands amd/intel/nvdia isn't enough

More competition won't work. They will only try match prices up against what Nvidia will charge. And if they offer something much better they will do as AMD did with Ryzen 5950X. First mainstream chips that reached $800 USD. 

 

There was no reason that AMD wanted MSRP $899 USD for 7900XT. A $699 Gtaphics cards. The competition don't works!!!

 

33 minutes ago, ryan said:

also @Papusan I was thrown back after watching the 3060 vs 4060 laptop comparison by jarrods tech, like a 10 percent bump for 300 bucks. its very frustrating seeing technology de-evolve or go into stagnation due to greed. we need more competition as it stands amd/intel/nvdia isn't enough

 

See.... Not so sure 4060Ti will be so enjoyable even with fake frame features.... 15% real performance uplift from the 2 years older 3060Ti with same amount money is a Joke. And this with +50% higher boost clocks.

 

A note on DLSS 3

Nvidia leans heavily on DLSS Frame Generation throughout its announcement to make the 4060 Ti look even faster than the 3060 Ti, which is all well and good. For games that support it, DLSS FG can take frames rendered by your GPU and generate an interpolated frame to show in between them, hypothetically boosting frame rates without asking the GPU to render more frames.

 

But there are three problems with DLSS FG, especially on mid-range GPUs. First, not all games support it, though many newer and more-demanding ones do (and pending Unreal Engine 5 support should make it even easier for developers to add support in new games). Second, it adds latency, which can make controls feel laggier in fast-paced shooters or other games where precision is important.

 

And third, it works best in situations where your GPU is already pushing a decent frame rate. More rendered frames means that Nvidia's AI algorithm needs to "guess" less when generating its interpolated frames, and imperfections in those AI-generated frames are harder to spot at 120 fps than they are at 40 fps. We encountered some of this in our RTX 4070 review, where more visual artifacts were more noticeable in games like Cyberpunk 2077 where the GPU was rendering at closer to 30 fps than 60.

 

Obviously, buying a midrange-to-low-end graphics card has always been about making compromises and balancing smooth frame rates with resolution and graphical fidelity. Despite its tradeoffs, DLSS FG is a powerful lever that people can pull to try to find a balance that works for them. It just isn't a substitute for better rasterized game performance, which benefits all games in all scenarios rather than just a subset.

[ryan]

supply and demand I suppose. If people are buying them and they are selling like hot cakes there is no reason to drop prices, I think US the consumer need to stop behaving like consumer cows and stop buying this crap. will it happen? probably not. I think the 4090 from what you gentlemen have shown is worth it's weight in gold, it really is a great card, but not everyone has 2000 for just a card. But there is just enough people buying them to keep this pattern alive

[Mr. Fox]

22 minutes ago, Papusan said:

4060Ti. What a fantastic generation performance uplift. 15 percent faster performance than the 3060 Ti in games without the DLSS Frame Generation feature enabled. Up to 52% higher boost clocks but only 15 percent faster real performance... Yep, reducing CUDA cores and MEM Bus vs its predessor come at a cost. Yep, what an improvement. Rather a Joke. 

Nvidia introduces $399 RTX 4060 Ti and $299 4060 without introducing a price hike arstechnica.com

20 minutes ago, ryan said:

also @Papusan I was thrown back after watching the 3060 vs 4060 laptop comparison by jarrods tech, like a 10 percent bump for 300 bucks. its very frustrating seeing technology de-evolve or go into stagnation due to greed. we need more competition as it stands amd/intel/nvdia isn't enough

What better way to sell your flagship that destroys every other option available by a huge margin--but, with a huge price tag attached to it--than to make all of the alternative options an undesirable, overpriced and lame Mickey Mouse excuse for an upgrade?

20 minutes ago, ryan said:

Epic games store actually has a pretty kickin game for free right now just caught it in the nick of time. death stranding is free and a great game to benchmark as it's well optimised and the performance to visuals trade off is one of the best.

That is a pretty decent game. I have played it and it's not bad. I am surprised to see it available for free on Epic. Thanks for the heads up. Anyone that doesn't have it should snag it while they can.

[cylix]

https://wccftech.com/amd-radeon-maintains-ten-week-long-gpu-sales-lead-over-nvidia-geforce-in-germany/

[Papusan]

LOOOL

 

In the most VRAM demanding games using ultra settings, like The Last of Us Part I, for example, the RTX 3060 may end up outperforming the newer 4060 due to its larger VRAM buffer.

 

Nvidia stated in their briefing with us that the RTX 4060 Ti isn't designed to be used on ultra settings; being a mid-range card, it's more suited to high-ish settings at modest resolutions. Yep, I'm 110% sure this new cards will be amazing in 720P forwards, HaHa

 

https://www.techspot.com/news/98731-nvidia-launches-three-new-geforce-rtx-4060-gpus.html

 

 

[ryan]

I saw the same thing. I was thinking should I sell and upgrade to a 4070 or 4060. haha funny thing is the difference is so small even at 200-300 bucks more is a rip off

[electrosoft]

1 hour ago, cylix said:

https://wccftech.com/amd-radeon-maintains-ten-week-long-gpu-sales-lead-over-nvidia-geforce-in-germany/

 

While AMD leads in volume by ~6%, Nvidia leads in revenue by 13% for the most current week along with an overall average sale price ~33% higher over the entire data chart.

 

Nvidia is taking the Apple vs Android approach choosing average sale price and profits over volume sales.

 

Still glad to see AMD make some ruckus and I suspect Nvidia is feeling it from AMD and an overall economic downturn world wide. Newest rumors has Nvidia putting the 4090ti on the back burner to focus on a Titan card. I wouldn't be surprised if they skipped the TI this generation or a very late cycle dump out with demand being lower overall even this early into the product launches. Only thing that remains sold out consistently are 4090 FEs.

 

4080 and 7900xtx neck and neck (170 vs 160)

4070 dominated everything in units sold.

 

 

 

 

[Ashtrix]

6 hours ago, Talon said:

 

Asus BIOS engineer showing Steve how to actually measure voltage. 

 

 

I was about to post this.,..saw it yest. Glad to see it.

 

All these Techtubers are just monetizing heavily on the AMD - ASUS situation to max and fooling people enmasse esp GN throwing a bunch of Electron Microscope zoom pics and look at comments everyone is just simply salivating and no one is really into trying to find the truth nor even think about the inconclusive outcome.

 

Basically It's known knowledge to all of us here. Socket Sense vs Die Sense.

 

ASUS APEX on Intel allows Die Sense readings from Z590 series exclusively using a hardware switch on mobo enabling that feature vlatch (bro @Papusan should know it..) and the ROG engineer in that above video knows how the AMD boards are made, basically unlike Intel there's no physical switch at-least on C8H, they just incorporated it onto the top end stack of AM5 boards so any HW sensor data polling will result in accurate reading (perhaps related to how Intel vs AMD Mobo design works) in our case HWInfo.

 

Now look at Igors lab, and HWbusters YT channel, both use same tool to monitor the Motherboard read out points and show it on the same UI claiming how the 0.05 voltage is extra over the AMD 1.3v fixed cap.  As such reputed testers fail to capture another data point. Both of them have Gigabyte board.. use same tool and miss same thing.

 

GB posted a video too check below highlighting how the HW voltage readouts are not accurate because of the Die Sense integration (they are dumb as rock though why not just speak like ASUS lol just muted and showing it normal person won't even understand even 1% let alone grasp all that)

 

 

 

It's really insane how they all are using a single data point. GN used the readout point from near PCIe slot if we see in their video  snapshot below..Yea it's all cool lol but why just completely ignore the HWinfo..

 

 

These lads failed to be called as Techtubers. Just another YT reviewer crowd capitalizing on the sensationalism.

 

Anandtech captured data on ASRock X670E, and they used HWInfo which is why their data is shown accurate but they also missed the point on Die Sense. However their testing is interesting as they test the total Power draw and Current plus Temperatures as well across all AGESA revisions.

 

Here's more... from the actual real deal in HW Builzdoid (he already stated how 12VPHWR did not have good ceiling vs 8Pin) - https://nitter.snopyta.org/Buildzoid1/status/1658131479275216896#m clearly mentions how the the testing is not accurate.

 

Second if you go all the way below that tweet responses, some fellow Nords of bro Papu checked it themselves too by following the ASUS engineer and found out same result.. They used HWInfo + ASUS engineer mentioned points to read actual Die Sense VSoC. AMD's new AGESA really locks it down is the conclusion. Even with increased LLC.

 

AMD has MCM design I was not aware of anything how VSoC is related to die sense, if it was Intel it would have struck me in a second on die sense, as we do only check VCore there's no MCM design to read another voltage point.

 

 

 

 

[Papusan]

 

 

 

Quote

Cheaper to make and will reach out to the bigger audience - Why lower TDP is preferable for low/mid end cards............

Here's why nvidia prefered lower TGP for their new gen low-mid end graphics cards. More people can get new and modern graphics in their mini pc's. Those that prefer bigger chassis with better cooling capacity for maximum performance from lower tier cards will lose as always. Nice. 

First Mini-ITX GeForce RTX 40 graphics card announced
https://videocardz.com/newz/first-mini-itx-geforce-rtx-40-graphics-card-announced

 

 

On 5/18/2023 at 6:32 AM, Mr. Fox said:

I do not believe the 12VHPWR connector was necessary and I do not consider it to be a real improvement.

 

The 12VHPWR connectors is the Gods gifts to the repair shops. The 4090 cards is very powerful and will have a long life in the used market with new owners. This means many cards will be out of warranty and the only fix will be by the repair shops. Yep, the 4090's is a Gold mine for any electronic repair shop.

 

I'm quite sure the 12VHPWR connectors will come in new revisions for 5000 series cards. For the 4000 series cards its too late... Yep, I think AMD will skip this new connector for current gen Radeon cards (possible 7950XTX). They can't afford bad publicity after their try knocking down nvidia due the smoking 12VHPWR connectors before 7000 series cards release. + the increased costs will make it even harder for AMD to do such expensive changes.

 

Here is 3 videos in a row from the last 5 days from NorthridgeFix. And I think he will make ready for the 4th or 5th video for this week or next (he got new ones in for repairs). And this delicate 12VHPWR mess won't stop by itself.... 

 

 

The last one is for @electrosoft🙂 From today. The weak point was the 90 degree CableMod adapter. Not the 12VHPWR adaper cable. Both fully plugged in. The 90 degree adapter plug was fully molded with the GPU connector. Stroger connection than with any of the best glue you can find on the market, LOOL

 

4090 FE Connector melted with Cablemod Adapter installed. 40 series cards should be recalled

On 5/18/2023 at 3:09 AM, electrosoft said:

 

Cablemod finally shipped my replacement out. It is coming directly from China.

 

 

Now I just have to set up sending in my AC LF 420 to Arctic for a new unit since mine is in the recall window. During the transition to AMD, swapping over to EVGA freed the monster 420. 🙂