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

[Papusan]

 

Our old friend F. Azor continue as he did in Dell Alienware. Not much have changed from that man. But he is  right in one thing...There is market for 8GB graphics cards trash. No in hell people should pay $300 and upwards for 8GB cards in 2025. 

AMD defends RX 9060 XT 8GB, says majority of gamers have no use for more VRAM

LOOL

 

And there is something called double standard (AMD in a nutshell). Or AMD have adopted the salesman F. Azors thinking about how to run the business.

 

AMD has been heavily critical of NVIDIA in the past for its memory limitations, but now that it has launched a Radeon RX 9060 XT with 8 GB , it has changed its mind. In recent years, AMD has repeatedly criticized NVIDIA for only using 8 GB of VRAM in its modern graphics cards. But now it is no longer of the same opinion as it has followed in its footsteps.

[tps3443]

7 minutes ago, Papusan said:

 

Our old friend F. Azor continue as he did in Dell Alienware. Not much have changed from that man. But he is  right in one thing...There is market for 8GB graphics cards trash. No in hell people should pay $300 and upwards for 8GB cards in 2025. 

AMD defends RX 9060 XT 8GB, says majority of gamers have no use for more VRAM

It’s very possible that the majority of PC gamers, install a game and simply play it. They won’t even change their default graphic settings. And they don’t even care. I was the one to teach my own son. But I remember when he’d install a brand-new game. He couldn’t care less about his graphics, or his frame rate. He would just be so excited to play the game that it didn’t matter. The 9060XT 8GB is the “On/Off Switch for PC gamers” 

 

Option A: No graphics card, (Cannot play games at all)

Option B: I have a 9060XT, I can actually play games now.

 

🙂

 

They wanna force an upgrade out of em down the road too! 😂
 

[Mr. Fox]

Hey Brother @jaybee83 have you tried any of this stuff yet? You get it straight from Splave out of Austin, TX.

 

https://splavepc.com/

 

Essentially the same performance as KPx and Duronaut. The KPx application had been in use a week or so. The Splave One and Duronaut were fresh applications. The Splave one did not work as well with a manual spread. I redid it with a line down the middle and let the waterblock spread it.

 

Bottom line: it doesn't matter which one you choose. Use what you have. If the Duronaut is actually more durable (as claimed) it may be the better long-term choice.

[Mr. Fox]

It is puzzling to me that anyone would pay $1,000 USD for an MSI Godlike or ASUS Extreme motherboard. Other than a couple of cheap to make gadgets and gimmicks they offer little or nothing and seem to be a total waste of money, especially on a Ryzen build. It applies to Intel as well. Just a scam to milk more money from those easily influenced by marketing rhetoric I think.

 

https://www.overclock.net/posts/29469837/

[Papusan]

19 hours ago, tps3443 said:

It’s very possible that the majority of PC gamers, install a game and simply play it. They won’t even change their default graphic settings. And they don’t even care. I was the one to teach my own son. But I remember when he’d install a brand-new game. He couldn’t care less about his graphics. He would just be so excited to play the game that it didn’t matter. The 9060XT 8GB is the “On/Off Switch for PC gamers” 

 

Option A: No graphics card, (Cannot play games at all)

Option B: I have a 9060XT, I can actually play games now.

 

They wanna force an upgrade out of em down the road too! 😂
 

 

Instead for buy overpriced low/mid end 8GB graphics cards 30-50% above MSRP there is options for those that don't care. 

 

Do Gamers Need A Graphics Card? iGPU Performance Explored

 

I remember the flack Nvidia got from launching two different 4080 SKUs. One of the 4080 become withdrawn due wrong naming. AMD does the same in 2025 for 9060 XT. They never learned from Nvidia's failure/scam. The 8GB cards should be branded non XT and re labeled as 9050 8GB Edition...

 

 

And nvidia continue with bad QC. 

Nvidia releases emergency RTX 5060-series firmware to fix blank screens on reboot

[jaybee83]

15 hours ago, Mr. Fox said:

Hey Brother @jaybee83 have you tried any of this stuff yet? You get it straight from Splave out of Austin, TX.

 

https://splavepc.com/

 

Essentially the same performance as KPx and Duronaut. The KPx application had been in use a week or so. The Splave One and Duronaut were fresh applications. The Splave one did not work as well with a manual spread. I redid it with a line down the middle and let the waterblock spread it.

 

Bottom line: it doesn't matter which one you choose. Use what you have. If the Duronaut is actually more durable (as claimed) it may be the better long-term choice.

 

huh! first time ive heard of that paste! seems to perform in the upper echelons, nice 🙂 so far im pretty happy with duronaut, especially the price point is without competition tbh. also still got a bunch of GC extreme and kryonaut, next to the gallinstan and conductonaut extreme i recently bought. for now im all pasted out 😄 

 

12 hours ago, Mr. Fox said:

It is puzzling to me that anyone would pay $1,000 USD for an MSI Godlike or ASUS Extreme motherboard. Other than a couple of cheap to make gadgets and gimmicks they offer little or nothing and seem to be a total waste of money, especially on a Ryzen build. It applies to Intel as well. Just a scam to milk more money from those easily influenced by marketing rhetoric I think.

 

https://www.overclock.net/posts/29469837/

 

well....as in most cases, "it depends" 😄 i for one am very happy about the crazy storage support that the extreme provides, without it i wouldnt have been able to fully stretch my SSD legs based on the high capacity needs that i have. like 99% of the other AM5 boards dont have these capabilities or have them with compromises (i.e. the need to dial down the primary gpu slot). 

but i agree that in 99% of use cases, godlike / extreme mobos are definitely not needed, especially if all one does is gaming!

 

alright boiz, the RAM kit binning continues. ive finished another two kits, the G. Skill DDR5-8400 and the Patriot DDR5-8000. Although both managed to pass 7600, theyre out of the race because 7800 errored out quite quickly in both cases. Thus far, the G.Skill 8200 kit is by far the strongest contendor, likely capable of getting 7800 stable with a few more adjustments.

 

We also have an additional contender now, bringing the total to 10(!) kits, including the Teamgroup kit im already sporting: turns out the shop i ordered from had the Trident Z5 Neo 32GB 8000 kit labeled incorrectly as the 48GB kit, so i guess i get to compare both the 2x24 and 2x16gb kits 😄 should be interesting after what bro @Mr. Fox mentioned!

 

edit: nevermind LOL. turns out im actually gonna end up with TWO 2x24GB DDR5-8000 Trident Z5 NEO kits 😄 heres the gist: i was using a price search engine to determine which kits and shops i would utilize. the link for the 2x24GB kit led to the shop's listing of the 2x16GB kit which i accidentally then ordered. but in the end they did send me the 2x24GB kit, after all 😄 after i send the order, i separately ordered the 2x24GB kit at another shop, so here we are: two identical kits hahaha. doesnt matter, ill still test them and see whats up 🙂 sorry for the confusion guys, it got a bit messed up 😛 

 

Notes:

Striked through: Out of the race

Blue: current champ

Percentage in parenthesis: how far did the stress test go before it errored out

 

8000 Mhz

G.Skill Trident Z5 RGB F5-8000J4048E24GX2-TZ5RK 1.30V (newer SKU)

G.Skill Trident Z5 RGB F5-8000J4048F24GX2-TZ5RK 1.35V (older SKU)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 1

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 2

Patriot Memory PVXR548G80C38K = 7600 / 7800 (3%)

Kingston Fury KF580C36RLAK2-48

Lexar ARES Gen2 RGB LD5U24G80C40BR-RGD = 7600 (20%)

 

8200 Mhz

G.Skill Trident Z5 RGB F5-8200J4052F24GX2-TZ5RK = 7600 / 7800 (97%)

 

8400 Mhz

G.Skill Trident Z5 RGB SP-8400J4052G24GX2-TZ5RK = 7600 / 7800 (1%)

[Mr. Fox]

21 minutes ago, jaybee83 said:

 

huh! first time ive heard of that paste! seems to perform in the upper echelons, nice 🙂 so far im pretty happy with duronaut, especially the price point is without competition tbh. also still got a bunch of GC extreme and kryonaut, next to the gallinstan and conductonaut extreme i recently bought. for now im all pasted out 😄 

 

 

well....as in most cases, "it depends" 😄 i for one am very happy about the crazy storage support that the extreme provides, without it i wouldnt have been able to fully stretch my SSD legs based on the high capacity needs that i have. like 99% of the other AM5 boards dont have these capabilities or have them with compromises (i.e. the need to dial down the primary gpu slot). 

but i agree that in 99% of use cases, godlike / extreme mobos are definitely not needed, especially if all one does is gaming!

 

alright boiz, the RAM kit binning continues. ive finished another two kits, the G. Skill DDR5-8400 and the Patriot DDR5-8000. Although both managed to pass 7600, theyre out of the race because 7800 errored out quite quickly in both cases. Thus far, the G.Skill 8200 kit is by far the strongest contendor, likely capably of getting 7800 stable with a few more adjustments.

 

We also have an additional contender now, bringing the total to 10(!) kits, including the Teamgroup kit im already sporting: turns out the shop i ordered from had the Trident Z5 Neo 32GB 8000 kit labeled incorrectly as the 48GB kit, so i guess i get to compare both the 2x24 and 2x16gb kits 😄 should be interesting after what bro @Mr. Fox mentioned!

 

edit: nevermind LOL. turns out im actually gonna end up with TWO 2x24GB DDR5-8000 Trident Z5 NEO kits 😄 heres the gist: i was using a price search engine to determine which kits and shops i would utilize. the link for the 2x24GB kit led to the shop's listing of the 2x16GB kit which i accidentally then ordered. but in the end they did send me the 2x24GB kit, after all 😄 after i send the order, i separately ordered the 2x24GB kit at another shop, so here we are: two identical kits hahaha. doesnt matter, ill still test them and see whats up 🙂 sorry for the confusion guys, it got a bit messed up 😛 

 

Notes:

Striked through: Out of the race

Blue: current champ

Percentage in parenthesis: how far did the stress test go before it errored out

 

8000 Mhz

G.Skill Trident Z5 RGB F5-8000J4048E24GX2-TZ5RK 1.30V (newer SKU)

G.Skill Trident Z5 RGB F5-8000J4048F24GX2-TZ5RK 1.35V (older SKU)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 1

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 2

Patriot Memory PVXR548G80C38K = 7600 / 7800 (3%)

Kingston Fury KF580C36RLAK2-48

Lexar ARES Gen2 RGB LD5U24G80C40BR-RGD = 7600 (20%)

 

8200 Mhz

G.Skill Trident Z5 RGB F5-8200J4052F24GX2-TZ5RK = 7600 / 7800 (97%)

 

8400 Mhz

G.Skill Trident Z5 RGB SP-8400J4052G24GX2-TZ5RK = 7600 / 7800 (1%)

I really think part of what you are finding is poor firmware support for the 24GB modules. EOL platforms (Intel and AMD) are never going to get the level of firmware support they deserve because the expectation is for everyone to waste money on a new motherboard. I think it is deliberate and orchestrated. The 24GB modules also do not work well on Intel unless the motherboard OEM gives it the effort needed. And, it usually takes two or three BIOS updates focused on that before it works well.

[jaybee83]

2 minutes ago, Mr. Fox said:

I really think part of what you are finding is poor firmware support for the 24GB modules. EOL platforms (Intel and AMD) are never going to get the level of firmware support they deserve because the expectation is for everyone to waste money on a new motherboard. I think it is deliberate and orchestrated. The 24GB modules also do not work well on Intel unless the motherboard OEM gives it the effort needed.

 

ouch, ure not gonna like this bro fox, the first Neo kit errored out right away at 7600. i reran the test and same result 😅

 

Notes:

Striked through: Out of the race

Blue: current champ

Percentage in parenthesis: how far did the stress test go before it errored out

 

8000 Mhz

G.Skill Trident Z5 RGB F5-8000J4048E24GX2-TZ5RK 1.30V (newer SKU)

G.Skill Trident Z5 RGB F5-8000J4048F24GX2-TZ5RK 1.35V (older SKU)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 1 = 7600 (1%)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 2

Patriot Memory PVXR548G80C38K = 7600 / 7800 (3%)

Kingston Fury KF580C36RLAK2-48

Lexar ARES Gen2 RGB LD5U24G80C40BR-RGD = 7600 (20%)

 

8200 Mhz

G.Skill Trident Z5 RGB F5-8200J4052F24GX2-TZ5RK = 7600 / 7800 (97%)

 

8400 Mhz

G.Skill Trident Z5 RGB SP-8400J4052G24GX2-TZ5RK = 7600 / 7800 (1%)

[Mr. Fox]

6 minutes ago, jaybee83 said:

 

ouch, ure not gonna like this bro fox, the first Neo kit errored out right away at 7600. i reran the test and same result 😅

 

Notes:

Striked through: Out of the race

Blue: current champ

Percentage in parenthesis: how far did the stress test go before it errored out

 

8000 Mhz

G.Skill Trident Z5 RGB F5-8000J4048E24GX2-TZ5RK 1.30V (newer SKU)

G.Skill Trident Z5 RGB F5-8000J4048F24GX2-TZ5RK 1.35V (older SKU)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 1 = 7600 (1%)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 2

Patriot Memory PVXR548G80C38K = 7600 / 7800 (3%)

Kingston Fury KF580C36RLAK2-48

Lexar ARES Gen2 RGB LD5U24G80C40BR-RGD = 7600 (20%)

 

8200 Mhz

G.Skill Trident Z5 RGB F5-8200J4052F24GX2-TZ5RK = 7600 / 7800 (97%)

 

8400 Mhz

G.Skill Trident Z5 RGB SP-8400J4052G24GX2-TZ5RK = 7600 / 7800 (1%)

16GB modules or 24GB? The 24GB modules suck on the Gene. They work perfectly on the generation newer Master. If you are testing only 24GB modules I am not surprised. If you want that you're probably going to need to get an X870E motherboard to find the appropriate attention to it at the firmware level. The Gene has a BIOS update intended to support 24GB modules, but using them still sucks in terms of performance and stability. I really do not believe it is a hardware issue, just lack of proper attention to firmware.

[jaybee83]

15 minutes ago, Mr. Fox said:

16GB modules or 24GB? The 24GB modules suck on the Gene. They work perfectly on the generation newer Master. If you are testing only 24GB modules I am not surprised. If you want that you're probably going to need to get an X870E motherboard to find the appropriate attention to it at the firmware level. The Gene has a BIOS update intended to support 24GB modules, but using them still sucks in terms of performance and stability. I really do not believe it is a hardware issue, just lack of proper attention to firmware.

 

all kits im testing are 24GB modules. and yes ofc, theres no hardware issue, everything works as intended. its likely optimizations on the firmware level as u mentioned, with some tubtimings tuned a bit differently on each kit.

 

i just saw that asus dropped a new bios with "optimizations for high storage RAM sticks" 😄 not sure if 24GB counts though, they likely mean 32GB+ modules.

[tps3443]

4 hours ago, jaybee83 said:

 

all kits im testing are 24GB modules. and yes ofc, theres no hardware issue, everything works as intended. its likely optimizations on the firmware level as u mentioned, with some tubtimings tuned a bit differently on each kit.

 

i just saw that asus dropped a new bios with "optimizations for high storage RAM sticks" 😄 not sure if 24GB counts though, they likely mean 32GB+ modules.

Are you planning to watercool your 5090?

[Reciever]

 

https://www.3dmark.com/3dm/134254982

 

Got #3

 

Dont think I'll be able to get much higher for Steel Nomad but #3 is pretty good for air cooling

 

[Mr. Fox]

19 minutes ago, Reciever said:

 

https://www.3dmark.com/3dm/134254982

 

Got #3

 

Dont think I'll be able to get much higher for Steel Nomad but #3 is pretty good for air cooling

 

Nice job!

[Reciever]

27 minutes ago, Mr. Fox said:

Nice job!

Thanks, spent the better part of the day working with my 7900 XTX trying to figure out its behavior. 

 

Finally came to terms with despite using an A/C unit freezing in my own room, the stock fans on the Red Devil were not enough to cool 800w (surprised they took me this far to be honest). Replaced with 3x92mmx45mm fans powered externally. That allowed me to get this score, along with the application of PTM7950.

 

The hotspot temperature is really the only thing that I am taming on this card, GPU temp might hit 62c with the fans maxed.

[electrosoft]

2 hours ago, Reciever said:

 

https://www.3dmark.com/3dm/134254982

 

Got #3

 

Dont think I'll be able to get much higher for Steel Nomad but #3 is pretty good for air cooling

 

 

44 minutes ago, Reciever said:

Thanks, spent the better part of the day working with my 7900 XTX trying to figure out its behavior. 

 

Finally came to terms with despite using an A/C unit freezing in my own room, the stock fans on the Red Devil were not enough to cool 800w (surprised they took me this far to be honest). Replaced with 3x92mmx45mm fans powered externally. That allowed me to get this score, along with the application of PTM7950.

 

The hotspot temperature is really the only thing that I am taming on this card, GPU temp might hit 62c with the fans maxed.

 

Very nice!  Top three on air....boom!

 

4 hours ago, tps3443 said:

Are you planning to watercool your 5090?

 

I'm pretty sure @jaybee83 doesn't watercool his cards. I think I asked him before and the answer was no.

 

 

 

 

 

[jaybee83]

12 hours ago, tps3443 said:

Are you planning to watercool your 5090?

 

7 hours ago, electrosoft said:

 

 

Very nice!  Top three on air....boom!

 

 

I'm pretty sure @jaybee83 doesn't watercool his cards. I think I asked him before and the answer was no.

 

 

 

 

 

 

correct, when it comes to gpus im pretty boring, air cooling only for me. thats why i went for the suprim to get the best air cooler 🙂

 

btw GOOD NEWS GAIZ! i hit 8000 STABLE on one of the kits yeeeehaw! so this is indeed a very interesting finding, were getting a spread of at least 600 MT/s between the worst and best performing kit on the same cpu + mobo. u basically never see any reviewers doing this, they either bin cpus (seldom) or mobos (mostly). 

so this is one of the g.skill 8000 kits, the one running at 1.30V for the XMP profile. i also tried 8100 and it booted but then hang during the windows login screen. dont matter tho, ill take 8000! 

so this leaves 3 kits still to be tested, two out of which are already on their way to me.

stay tuned!

 

here the current standings:

 

8000 Mhz

G.Skill Trident Z5 RGB F5-8000J4048E24GX2-TZ5RK 1.30V (newer SKU) = 7600 / 7800 / 8000 / 8100 (Windows Login)

G.Skill Trident Z5 RGB F5-8000J4048F24GX2-TZ5RK 1.35V (older SKU)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 1 = 7600 (1%)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 2

Patriot Memory PVXR548G80C38K = 7600 / 7800 (3%)

Kingston Fury KF580C36RLAK2-48

Lexar ARES Gen2 RGB LD5U24G80C40BR-RGD = 7600 (20%)

 

8200 Mhz

G.Skill Trident Z5 RGB F5-8200J4052F24GX2-TZ5RK = 7600 / 7800 (97%)

 

8400 Mhz

G.Skill Trident Z5 RGB SP-8400J4052G24GX2-TZ5RK = 7600 / 7800 (1%)

[electrosoft]

9 hours ago, jaybee83 said:

 

 

correct, when it comes to gpus im pretty boring, air cooling only for me. thats why i went for the suprim to get the best air cooler 🙂

 

btw GOOD NEWS GAIZ! i hit 8000 STABLE on one of the kits yeeeehaw! so this is indeed a very interesting finding, were getting a spread of at least 600 MT/s between the worst and best performing kit on the same cpu + mobo. u basically never see any reviewers doing this, they either bin cpus (seldom) or mobos (mostly). 

so this is one of the g.skill 8000 kits, the one running at 1.30V for the XMP profile. i also tried 8100 and it booted but then hang during the windows login screen. dont matter tho, ill take 8000! 

so this leaves 3 kits still to be tested, two out of which are already on their way to me.

stay tuned!

 

here the current standings:

 

8000 Mhz

G.Skill Trident Z5 RGB F5-8000J4048E24GX2-TZ5RK 1.30V (newer SKU) = 7600 / 7800 / 8000 / 8100 (Windows Login)

G.Skill Trident Z5 RGB F5-8000J4048F24GX2-TZ5RK 1.35V (older SKU)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 1 = 7600 (1%)

G.Skill Trident Z5 NEO RGB F5-8000J4048G24GX2-TZ5NR Kit 2

Patriot Memory PVXR548G80C38K = 7600 / 7800 (3%)

Kingston Fury KF580C36RLAK2-48

Lexar ARES Gen2 RGB LD5U24G80C40BR-RGD = 7600 (20%)

 

8200 Mhz

G.Skill Trident Z5 RGB F5-8200J4052F24GX2-TZ5RK = 7600 / 7800 (97%)

 

8400 Mhz

G.Skill Trident Z5 RGB SP-8400J4052G24GX2-TZ5RK = 7600 / 7800 (1%)

 

Ditto with my main rig card, but you water cool the CPU or AIO? You know what, just slap up a refresher pic of your rig.... 🤣

 

w00t! on the 8000! on a 2DPC x670e board, that is a massive win!