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

[Mr. Fox]

3 hours ago, Etern4l said:

 

Yes, I missed the fact that the "scales" are real are not just a cool design painted over lol. Clever.  How difficult are those to install on say G.Skill Trident Z5 or Corsair Vengeance? Warranty voided after removing the original heat spreader I guess?

Using force to remove the stock covers can be very dangerous in terms of tearing the chips off of the PCB if not done in a very specific manner. You have to wedge an object between the IC and the cover to push the cover off away from the IC rather than pulling on the cover externally. The first couple of times I used this method, I used a plastic pick tool that was longer than the module and very slowly and gently pushed it between the IC chips and the outside cover lengthwise a few millimeters at a time. The tension between the chips and the cover pushing them apart separated them. But this method was destructive to the adhesive pads used to glue the covers in place.

 

The safest way to get them off is to soak them in a shallow bath of lacquer thinner or acetone long enough to soften the adhesive on the pads and then they can be lifted off with minimal effort. I have used this method with two kits and it is a much better way. If you take care not to submerge the portion of the memory module that has the identification label on it, then it will be impossible to tell visually that anything has happened.

 

After the solvent has dried there is generally most or all of the adhesive left intact and you can reinstall the covers and it will look as if they have never been tampered with. You need to store the covers inside of the original plastic clamshell to avoid any dust getting on the adhesive over time so that they remain sticky in case you ever wish to sell or RMA the modules. Then you can put them back on again later and they'll look as good as new.

 

If no harm is done removing the covers there is absolutely no legitimate or honorable basis for the warranty to be affected. You would actually be doing the manufacturer a huge favor and enhancing the life and stability of the memory modules by getting those crappy heating blankets off of them. Knowing that most of the OEMs look for lame excuses to not stand behind their products, it is wise to take steps to avoid giving them something to leverage as an excuse for shirking their responsibility for selling a good quality, long-lasting product.

[Etern4l]

Cool, good advice as always. I look forward to bro @electrosoft's report on this.

[Papusan]

The Lightning have landed here home😎 One of the very few unicorn in the graphics cards world. Say hello to the 11 years old MSI N680GTX Lightning. Always wanted me a Lightning card (fully unlocked). This graphics card model together with K|ngp|n, HOF and the Asus Matrix cards have always been the top of the crop for high performance graphics cards. Here are some pict of the old beauty for the pict lover @electrosoft Enjoy🙂 

 

No plastic fantastic as new and modern. Pure metal all over. Who said newer has to be better, LOOL

 

Yep, the card is the full blown Lighning. Not the gimped down N680GTX Lighning-L

 

 

 

 

 

 

 

There is no guarantee you'll get the best top bin (asic) even with the best models. So nothing different than today's modern best binned graphics cards. Will be fun see what love Msi gave this particular Lighning card. 
 

Zotac looks weird in comparison. The pipes is placed weird.

[ryan]

the team

AMEETTH

 

brilliant just caught that, i'll keep out of the thread until im versed but color me intrigued. your either a genius or a closet genius.

[Papusan]

I have put in the card. Barely space for the GPU- Reactor box to fit in below the Apex SSD heatsink. So I don't need to remove the Grey/blue Reactor cover. It would have been impossible to mount it correctly if the cover had been tiny bit larger🙂

 

And it seems it's alive. Both fans spins and the GPU is recognized in the OS. So a good start for the "new" old🙂

 

 

Broken games and broken prices for graphics cards. Just the way modern tech works nowadays. Yep, old is gold bro @ryan. I hope more PC gamers will migrate Roll Over over to consoles so they can kick Mr Jensen in the balls. Yep, AMD will get a gain from this but they have to follow Nvidia if they want to compete for Pc gaming.

 

 

Broken games and exorbitant GPUs: PC gaming is rough right now

https://www.pcworld.com/article/1805943/pc-gaming-is-rough-right-now-and-its-depressing.html

 

 

Who will break their bank and buy brand new 8GB mid-range graphics cards in middle 2023? Will new modern fake frame boost features save this cards from being thrown in the bin within next 2-3 years? And will game developers help Nvidia/AMD to get rid of their brand new 8GB graphics cards?

 

Gigabyte EEC Listing Confirms AMD RX 7600, Nvidia RTX 4060 Ti Cards for Launch
https://www.tomshardware.com/news/gigabyte-eec-listing-confirms-amd-rx-7600-nvidia-rtx-4060-ti-cards-for-launch

[Mr. Fox]

1 hour ago, ryan said:

the team

AMEETTH

 

brilliant just caught that, i'll keep out of the thread until im versed but color me intrigued. your either a genius or a closet genius.

???

 

What post are you responding to, Brother @ryan?

[ryan]

wrong thread. eternals thread and addy in ai thread

[Papusan]

The "new" old - MSI GTX 680 Lightning

 

https://hwbot.org/submission/5267104_papusan_3dmark_vantage___performance_geforce_gtx_680_62217_marks?recalculate=true

 

https://hwbot.org/submission/5267098_papusan_3dmark_vantage___extreme_geforce_gtx_680_27553_marks/

 

https://hwbot.org/submission/5267122_papusan_3dmark2001_se_geforce_gtx_680_188391_marks?recalculate=true

 

https://hwbot.org/submission/5267147_papusan_3dmark___fire_strike_geforce_gtx_680_9526_marks?recalculate=true

 

And right in front of Mr. Oviz in 3DM FS GPU score😀

https://hwbot.org/submission/5267149_papusan_3dmark___fire_strike_(gpu)_geforce_gtx_680_9977_marks?recalculate=true

[ryan]

https://www.3dmark.com/search#advanced?test=fs P&cpuId=&gpuId=747&gpuCount=1&gpuType=ALL&deviceType=ALL&storageModel=ALL&memoryChannels=0&country=&scoreType=overallScore&hofMode=false&showInvalidResults=false&freeParams=&minGpuCoreClock=&maxGpuCoreClock=&minGpuMemClock=&maxGpuMemClock=&minCpuClock=&maxCpuClock=

 

https://www.3dmark.com/fs/346850

 

I used to be numero uno. number 1 in the world with every card in a laptop. now Im having issues making average lol. ah times change, times change

 

I won't lie alot of it is luck(lottery) sometimes you get poorly binned cards I just fluked out like 4-6 times in a row. stopped around 680m

[Papusan]

MSI GTX 680 Lightning

 

https://hwbot.org/submission/5267210_papusan_aquamark_geforce_gtx_680_579521_marks?recalculate=true

 

https://hwbot.org/submission/5267228_papusan_3dmark___cloud_gate_geforce_gtx_680_51180_marks?recalculate=true

 

https://hwbot.org/submission/5267158_papusan_3dmark___fire_strike_extreme_geforce_gtx_680_4490_marks?recalculate=true

 

https://hwbot.org/submission/5267175_papusan_3dmark11___performance_geforce_gtx_680_14548_marks?recalculate=true

[ryan]

and this is my tweak that no one knows about. no one uses it and it can have a biggish impact on cpu performance unlike many many tweaks, this my bros is the real deal. ill tell you once and if you don't do it its your own fault that you didnt get the highest score/highest cpu performance. like this everyone including @johnksss @Papusan @Mr. Fox @Papusan @tps3443 @electrosoft and pretty much everyone.

 

here is before the tweak 0 changes

after tweak completely identical settings

 

the tweak

 

 

basically the second level data cache is set to off or dynamic, you search google to find your cpus mb of 2nd level data cache if say its 5 then thats 5x1024 you multiply it by 1024 and then in reg edit search secondleveldata cache(if its set to 0 you will get a boost) then enter the amount 5x1024 or 6x1024 or whatever yours is and you should see a boost thats my secret, and why i was always numero uno with old cpus/gpus and it still works to this day I literally just used it on my hp 320 dollar laptop and saw a 10 percent boost doing a 10 second tweak

 

 

[Papusan]

The new Ryzen burnout fixing AMD AGESA 1.0.0.7 BIOS is reportedly buggy tweaktown.com | May 5, 2023 

Recently we reported on a new and mandatory AMD AGESA 1.0.0.7 BIOS Firmware for all AM5 motherboards built for the new Ryzen 7000 CPUs. This a direct response to the burnout issues found on some Ryzen 7000X3D chips that were damaging both the chip and motherboard.

AMD's root cause for the issue was around voltage. The new firmware limits the SoC voltage to 1.3V. It adds other measures/fixes to ensure processors don't exceed their temperature specifications - while not affecting memory overclocking via AMD EXPO profiles.

 

With potentially damaged hardware on the line, no doubt there's a rush to get this out to as many customers (and board partners) as possible, but with ASUS using a BETA release of the new AGESA 1.0.0.7 BIOS firmware, it looks like there are still issues to work out, which could delay the proper and final rollout of firmware to fix the burnout issue by weeks.

 

So then, what's the current hold-up? Well, it seems that memory speeds are being affected. A report from leaker chi11eddog on Twitter notes that memory speeds are being cut down to DDR5-4400 compared to the DDR5-6000 speeds with the older AGESA 1.0.0.6 BIOS.

 

And what bugs will be baked in next AGESA BIOS Firmware? Hmmm, maybe the older AGESA firmware with manual tuning would be better?

 

[Papusan]

MSI GTX 680 Lightning

 

https://hwbot.org/submission/5268006_papusan_3dmark06_geforce_gtx_680_76289_marks?recalculate=true

 

https://hwbot.org/submission/5267949_papusan_3dmark05_geforce_gtx_680_100104_marks?recalculate=true

 

https://hwbot.org/submission/5268012_papusan_unigine_heaven___xtreme_geforce_gtx_680_2964.73_dx11_marks?recalculate=true

 

 

https://hwbot.org/submission/5268043_papusan_unigine_heaven___basic_geforce_gtx_680_5897.08_dx9_marks?recalculate=true

[Raiderman]

Ok all you lighting experts, I am getting fed up with this crap. Installed the new Water cooling set up, and the biggest issue I have is the stupid Argb hub. Everything Ive tried, and cannot get ports 2,3,4,5 to light anything up. Not to mention the PC industry created one of the worst header plugs for these stupid things. Does anyone know why the other ports wont light anything up?

 

[Mr. Fox]

1 hour ago, Raiderman said:

Ok all you lighting experts, I am getting fed up with this crap. Installed the new Water cooling set up, and the biggest issue I have is the stupid Argb hub. Everything Ive tried, and cannot get ports 2,3,4,5 to light anything up. Not to mention the PC industry created one of the worst header plugs for these stupid things. Does anyone know why the other ports wont light anything up?

 

The loop looks nice, bro.

 

Without knowing more details, my first impression is the hub is defective if only one port works.

 

Are the ports that are not working on an aftermarket external hub or the ports on the motherboard itself? Are they 3-pin 5V or 4-pin 12V?

 

Generally on an aftermarket external hub all of the ports are active. There are two types. The good type require no motherboard ports and have their own RGB controller. If you're referring to ports on the motherboard itself sometimes you have to enable them in the bios. ASUS crap requires installation of their bloatware trash to use the onboard lighting and Aura rarely works correctly. More often than not Aura is broken and malfunctions.

[Raiderman]

2 minutes ago, Mr. Fox said:

The loop looks nice, bro.

 

Without knowing more details, my first impression is the hub is defective if only one port works.

 

Are the parts that are not working on an aftermarket external hub or the ports on the motherboard itself? Are they 3-pin 5V or 4-pin 12V?

 

Generally on an aftermarket external hub all of the ports are active. There are two types. The good type require no motherboard ports and have their own RGB controller. If you're referring to reports on the motherboard itself sometimes you have to enable them in the bios. ASUS crap requires installation of their bloatware trash and it rarely works. More often than not it is broken and malfunctions.

It a Raijintek Hub, and came with the kit. It can control with or without connecting to the MB. It has 5 5v ports on one side, and 5 PWM connections on the other side.

[Mr. Fox]

12 minutes ago, Raiderman said:

It a Raijintek Hub, and came with the kit. It can control with or without connecting to the MB. It has 5 5v ports on one side, and 5 PWM connections on the other side.

Sounds as if the hub itself is bad. If you don't feel like messing with getting it replaced I think you can get a one to three splitter on Amazon and connect all three of your ARGB fan cables to the same port. At least it is the ARGB and not the fan ports that are not working.

 

Raijintek will probably send you a replacement hub. Let us know how the communications with them goes. I have never purchased anything from them before, but they have some nice products.

[Raiderman]

13 minutes ago, Mr. Fox said:

Sounds as if the hub itself is bad. If you don't feel like messing with getting it replaced I think you can get a one to three splitter on Amazon and connect all three of your ARGB fan cables to the same port. At least it is the ARGB and not the fan ports that are not working.

 

Raijintek will probably send you a replacement hub. Let us know how the communications with them goes. I have never purchased anything from them before, but they have some nice products.

The problem is all these companies use proprietary plugs to plug into their hub. They are not the standard 3 pin as you find on a MB header. I have 2, three pin MB headers, and 2 splitters, but cannot plug anything that came from Raijintek into them. Although they may plug in anyway?? I will see if they will. Thanks bro!

[electrosoft]

1 hour ago, Raiderman said:

It a Raijintek Hub, and came with the kit. It can control with or without connecting to the MB. It has 5 5v ports on one side, and 5 PWM connections on the other side.

 

I'm sure it's not the problem but I've encountered a few cheap/knock off hubs that wire their pseudo-aRGB ports in serial and require a specific order to light up.

 

Have you tried just plugging one fan header into each individual port (including the one that works) to see what happens?  How many ports total?

 

I'm using a thermalright hub in mine they sent me to test and it is as generic/real aRGB headers as it gets and just works and only works via the MB.

 

I do see value in those "proprietary" type hubs as you usually get a killer price for a hub and 4-6 fans. Having evaluated several of them I noticed the non aRGB "proprietary" connector is all the same and the fans are interchangeable.

 

 

 

 

[Mr. Fox]

Phanteks has used a goofy connector for RGB as long as I can remember and I have purchased adapters to convert to generic industry standard connectors. Hubs with a built-in controller are the best. Avoiding anything that relies on motherboard lighting is a better approach.

[Raiderman]

So I have submitted a warranty request through Newegg, but I highly doubt that it will be a very good outcome. They do not service their warranties directly, and rely on the reseller. I want new parts sent to me, before I return the old ones, because I do not want my system down for 2 to 3 weeks. I submitted a claim because the cpu block has a faulty led, and is super annoying. I dont notice it much using white as the color choice, but changing it to anything else results in undesired effects.

The picture does not show it well, but the brighter area is actually blue.

 

Edit: I was actually able to plug in those proprietary plugs onto my splitters to get everything working. Look at this mess

[Papusan]

 

A bit too late for 3070 and 3080@10GB cards. But will coming vram feature save the vram castrated 4060 and 4060Ti Nvidia graphics cards from running out of vram? We have now features as DLSS to boost performance... Next level is to use compression features to sell overpriced cards with bare minimum of needed vram. Yep, Nvidia know how to make money. 

 

We'll learn more about NTC and how NVIDIA plans to develop this tech at SIGGRAPH 2023 in August. It won't be riding to the rescue of current-gen GPUs, though, that's for sure - and it might be some way down the line before we see it realized. Mind you, Wronski sounds confident enough that it's maybe not as far off as we think...

 

NVIDIA Says It Spends 20% of Its Time on Hardware, 80% on software so they can further cut costs on HW level

 

NVIDIA's new texture compression tech could be the end of VRAM worries tweaktown.com | Gaming | May 5, 2023

NVIDIA has developed a new compression algorithm for textures that looks very impressive.

VideoCardz spotted the paper from NVIDIA, which is called 'Random-Access Neural Compression of Material Textures,' and was also shared on Twitter by Bart Wronski, a researcher at Team Green who's part of the project.

The Neural Texture Compression (NTC) aims to tackle the spiraling requirements for high-res textures, allowing them to better fit into the available RAM. We're told that it delivers 16x more texels than Block Compression or BC (standard GPU-based texture compression).

 

 

 

[Etern4l]

1 hour ago, Papusan said:

 

A bit too late for 3070 and 3080@10GB cards. But will coming vram feature save the vram castrated 4060 and 4060Ti Nvidia graphics cards from running out of vram?

 

The catch is that is that this will put additional load on on the GPU cores, so will compete for the compute capacity with other gaming engine functions. This also incurs 2.5-4x higher latency compared to the very fast dedicated texture decompression hardware (on a 4090), so a fast GPU will be required. Bye bye Turing and Ampere, upgrade to 4000/5000-class GPUs required. Brilliant!  

[ryan]

1 hour ago, Etern4l said:

This also incurs 2.5-4x higher latency (on a 4090), so a fast GPU will be required.

im searching for this source, where did you read that. my internet is castrated its not showing every link. Im really interested in this tech as I have a 6gb 3060. will we all see this or just 5k series

 

back. just found this not much else

 

https://www.tomshardware.com/news/nvidia-uses-neural-network-for-innovative-texture-compression-method

 

Practical Advantages and Disadvantages

Indeed, images demonstrated by Nvidia clearly show that NTC is better than traditional Block Coding-based technologies. However, Nvidia admits that its method is slower than traditional methods (it took a GPU 1.15 ms to render a 4K image with NTC textures and 0.49 ms to render a 4K image with BC textures), but it provides 16x more texels albeit with stochastic filtering. 

 

Nvidia claims that NTC textures are decompressed using matrix-multiplication hardware such as tensor cores operating in a SIMD-cooperative manner, which means that the new technology does not require any special purpose hardware and can be used on virtually all modern Nvidia GPUs.

 

https://www.guru3d.com/news-story/nvidia-introduces-revolutionary-neural-texture-compression-for-material-textures.html

 

NVIDIA researchers are set to reveal further details about this groundbreaking technology at SIGGRAPH 2023 on August 6. According to their paper, neural textures can be rendered in real-time, achieving up to 16x more texel than the BC methodology. The expense of a 4K render using NTC is 1.15ms, marginally higher than the 0.49ms cost observed on the RTX 4090.

 

1ms extra is really small, and would not be noticeable. I game on geforce now from time to time and when the lag is under 70-50ms I don't feel it, even when jerking mouse around. also it should be noted is 16 times more efficient. so for the same image quality the burden is smaller, the added power required is only if the actual texture is 16 times more detailed(higher resolution). edit. ok maybe not true, It says 4k render is more than double the latency which i think is misleading as why say its revolutionary if for the same image quality its harder on the gpu, i think its mis-worded.

 

 

TLDR:

 

Old 0.49ms

New 0.078ms(1.15/16-for same IQ)

 

less vram needed

 

Pain in the butt for programmers

[Etern4l]

1 hour ago, ryan said:

im searching for this source, where did you read that.

 

I just looked through the paper, it's there:

 

Table 4. Decompression performance for a 4k material texture set (Paving Stones). 

 

BC High NTC 0.2 NTC 0.5 NTC 1.0 NTC 2.25

0.49 ms 1.15 ms 1.46 ms 1.33 ms 1.92 ms

 

This is just a simple performance test rendering a single shape at 4K:

 

6.5.2 Decompression. We evaluate real-time performance of our method by rendering a full-screen quad at 3840 × 2160 resolution textured with the Paving Stone set, which has 8 4k channels: diffuse albedo, normals, roughness, and ambient occlusion. The quad is lit by a directional light and shaded using a physically-based BRDF model [10] based on the Trowbridge–Reitz (GGX) microfacet distribution [76]. Results in Table 4 indicate that rendering with NTC via stochastic filtering (see Section 5.3) costs between 1.15 ms and 1.92 ms on a NVIDIA RTX 4090, while the cost decreases to 0.49 ms with traditional trilinear filtered BC7 textures.

 

They call the performance "similar", but in fact the latency it's up to 4x higher and takes up CUDA/RT cores, elsewhere in the paper they state it "might be possible to partially hide this overhead", given a fast enough GPU... 

 

When ascertaining if this will be of benefit in a particular game for you, look at HWInfo while gaming. You can see both memory controller and GPU cores utilisation. My bet is that usually you will be bottlenecked on the GPU anyway, so yes - with this you may be able to see higher resolution textures, but at some cost to the FPS, which presumably doesn't look brilliant on a 6GB 3060 mobile to be begin with.