Nvidia Turing Product Reviews and Previews: (Super, TI, 2080, 2070, 2060, 1660, etc)

lol Steve is rather bored with Nvidia at the moment, worth a watch just for laughs. Granted there's not really anything to talk about with 2080 Super, it's slightly faster and as Steve put it, "phenomenally boring"

 
RTX 2080 Super Reviews 1
Gamer ZOTAC RTX 2080 SUPER
AnandTech RTX 2080 SUPER Founders Edition
AusGamers RTX 2080 SUPER Founders Edition
Bit-Tech (VIDEO) RTX 2080 SUPER Founders Edition
Bitwit RTX 2080 SUPER Founders Edition
BPS Customs RTX 2080 SUPER Founders Edition
ComptoirHardware RTX 2080 SUPER Founders Edition
ComputerBase RTX 2080 SUPER Founders Edition
Cowcotland RTX 2080 SUPER Founders Edition
ElChapuzasInformatico RTX 2080 SUPER Founders Edition
Eteknix (VIDEO) RTX 2080 SUPER Founders Edition
GIGABYTE RTX 2080 SUPER GAMING OC
PALIT RTX 2080 SUPER GameRock
ZOTAC RTX 2080 SUPER AMP Extreme
Eurogamer RTX 2080 SUPER Founders Edition
Forbes RTX 2080 SUPER Founders Edition
Golem RTX 2080 SUPER Founders Edition
Guru3D RTX 2080 SUPER Founders Edition
PALIT RTX 2080 SUPER GameRock LE
MSI RTX 2080 SUPER GAMING X Trio
HardwareBattle RTX 2080 SUPER Founders Edition
Hardware Canucks RTX 2080 SUPER Founders Edition
Hardware.info RTX 2080 SUPER Founders Edition
HardwareUnboxed RTX 2080 SUPER Founders Edition
HardwareLuxx RTX 2080 SUPER Founders Edition
HD Tecnologia RTX 2080 SUPER Founders Edition
Hot Hardware RTX 2080 SUPER Founders Edition
Hexus RTX 2080 SUPER Founders Edition
Igor’sLAB RTX 2080 SUPER Founders Edition
KitGuru (VIDEO) RTX 2080 SUPER Founders Edition
Lab501 GIGABYTE RTX 2080 SUPER GAMING OC
LegitReviews RTX 2080 SUPER Founders Edition
Optimum Tech RTX 2080 SUPER Founders Edition
Overclock3D RTX 2080 SUPER Founders Edition
OverclockersClub RTX 2080 SUPER Founders Edition
Paul’s Hardware RTX 2080 SUPER Founders Edition
PCGamesHardware (VIDEO) RTX 2080 SUPER Founders Edition
PCGamesN RTX 2080 SUPER Founders Edition
PCMasterRace LA RTX 2080 SUPER Founders Edition
PCPerspective RTX 2080 SUPER Founders Edition
TechPowerUP RTX 2080 SUPER Founders Edition
ZOTAC RTX 2080 SUPER AMP EXTREME
 
Brazilians make a 'RTX 2080 Ti Super' by fitting it with faster GDDR6
So here's what three musketeers in Brazil did, they removed 16 Gbps GDDR6 from a Galax RTX 2080 Ti Hall of Fame, and then planted that same GDDR6 memory on a GeForce GTX 2080 Ti which normally has 14 Gbps GDDR6. Surprisingly enough, that actually worked.

Obviously I need to make a few remarks here, there is no fully enabled GPU, ergo the shader processor count is the same (which would be different for a Super model). Secondly, there's not one 2080 Ti card we tested, that could not get its memory tweaked at 16 Gbps. In fact the MSI lighting we clocked at 16,548 MHz (effective data-rate).

But .. they 16 Gbps GDDR6 memory, can be tweaked as well, and they ended at 17,200 MHz and that is 3,200 MHz more than the memory that is on a standard 2080 Ti. An higher frequency was not possible on that memory because the memory controller would not allow this, according to the modders.
https://www.guru3d.com/news-story/brazilians-fit-rtx-2080-ti-super-with-faster-gddr6.html
 
NVIDIA Seemingly Dropping Price on GeForce RTX 2060 Cards To $299 To Tackle AMD’s Radeon RX 5600 XT
During our visit to the EVGA booth at CES 2020, we came by an interesting NVIDIA GeForce RTX 2060 card with an even more interesting price range. This card is the EVGA GeForce RTX 2060 KO Edition and it will cost just $299 US compared to the current RTX 2060 pricing of $349 US.
...
At $299 US, the new RTX 2060 could offer much better value over the Radeon RX 5600 XT even while being $20 US more expensive. AMD has shown the RX 5600 XT to be up to 15% faster than a GTX 1660 SUPER and 20% faster than a GTX 1660 Ti.

The NVIDIA GeForce RTX 2060, on the other hand, is up to 20% faster than the GeForce GTX 1660 SUPER while offering added features such as RTX, DLSS, etc. It would be interesting to see how the Radeon RX 5600 XT and the price-cut GeForce RTX 2060 perform against each other in the coming weeks.
https://wccftech.com/nvidia-geforce-rtx-2060-price-cut-299-usd-tackle-amd-rx-5600-xt/
 
Professional app results for these cards ($300) are pretty amazing as they have a TU-104 (BTX 2070 Super) instead of TU-106 die.
 
Last edited:
looks like 2060 ko's are 2070's that failed some tests, (edit: 2080's that failed)
The interesting times will happen if nvidia run out of failed chips
will they secretly ship fully working 2070 gpu's flashed as 2060 ko's
 
looks like 2060 ko's are 2070's that failed some tests, (edit: 2080's that failed)
The interesting times will happen if nvidia run out of failed chips
will they secretly ship fully working 2070 gpu's flashed as 2060 ko's
"KO" is not an Nvidia version, it's something only that specific vendor is doing. Other vendors are simply buying them and releasing as regular 2060's, which is what Nvidia intended. If Nvidia runs out then that specific vendor won't have anymore "KO" versions to sell, that's all. It's not a real product.
 
Review: Nvidia GeForce RTX 2080 Ti
February 6, 2020
As well as seeing how the RTX 2080 Ti performs with a representative range of DCC applications, I will be testing how far Nvidia’s RTX hardware can accelerate GPU rendering, and attempting to answer a couple of questions that readers posed in response to the previous review: how important is memory capacity when choosing a GPU for production work, and how important is it to use Nvidia’s Studio GPU drivers?
...
What is interesting to see is how performance increases when OptiX is enabled, and the software can offload ray tracing calculations to the RT cores of the RTX cards. In Redshift, the impact is relatively small – although bear in mind that version 3.0 is still in early access – but in the V-Ray benchmark, performance increases by 33-35%, and in the Blender benchmark, by 85-105%.
http://www.cgchannel.com/2020/02/review-nvidia-geforce-rtx-2080-ti/
 
GPU Compute render perf review with 20 GPUs
When it comes to content creation (render) titles, sometimes applications will automatically opt what's they deem is best, others even trigger extended more advanced APIs, Blender, for example, can now also make use of Optix opposed to CUDA, but only if you have a GeForce RTX based graphics card. OptiX is part of Nvidia GameWorks and a high-level (close to the hardware) API, meaning that it is designed to encapsulate the entire algorithm of which ray tracing is a part, not just the ray tracing itself. This is meant to allow the OptiX engine to execute the larger algorithm with great flexibility without application-side changes. We can only assume that Optix triggers and makes use of RT (Raytracing) cores on the GeForce RTX series graphics cards. So that's three Compute APIs available.

Also let me state, that it's not one or the other; Hybrid Rendering is what is applied the most: (running OpenCL or CUDA both the GPU and CPU) – GPU OpenGL/CUDA rendering can be performed on CPUs and GPUs at the same time allowing the Compute code to combine your CPUs and GPUs to utilize all available resources. For this article, we measure the effect of the GPU only.

In this article, we'll test the three content rendering titles with twenty graphics cards, and see what that brings us in performance. We do so with the three forenamed applications as well as the three available Compute APIs these applications can utilize.
https://www.guru3d.com/articles_pages/gpu_compute_performance_review_with_20_graphics_cards,1.html
 
looks like 2060 ko's are 2070's that failed some tests, (edit: 2080's that failed)
The interesting times will happen if nvidia run out of failed chips
will they secretly ship fully working 2070 gpu's flashed as 2060 ko's

That will be AF (I mean like ryzen 1600AF) :p
 
Unsupported: How to Make Dual NVLink Work on Windows 10
In light of that experience, I've taken time in the last couple of weeks to go through and test several different NVIDIA drivers on two configurations: four GeForce RTX 2080 Ti cards as well as four Quadro RTX 6000s, both set up in two physically bridged pairs. In this article I will chronicle what I found worked, what didn't work or behaved oddly, and where we are at with this as a company as a result.
...
Can you make two pairs of NVLink cards work in Windows 10? Yes, by using older drivers... but Windows may update them at any time, potentially messing up the configuration. You also wouldn't be able to utilize improved performance or added features from newer drivers, and there is no guarantee that future versions of Windows will work with the older drivers. All in all, not a great solution.
https://www.pugetsystems.com/labs/a...-to-Make-Dual-NVLink-Work-on-Windows-10-1688/
 
Nvidia released RTX Voice Beta

When you're working, playing, and creating content from home, you want to keep background noise to a minimum. NVIDIA RTX Voice is a new plugin that leverages NVIDIA RTX GPUs and their AI capabilities to remove distracting background noise from your broadcasts and voice chats.
RTX Voice allows users to:
Go live or attend a meeting remotely without worrying about finding a quiet place. Suppress background noise from players in loud environments, making incoming audio easier to understand.

https://www.nvidia.com/en-us/geforce/forums/broadcasting/18/361740/rtx-voice-beta/

 
RTX 2060 at 2040/8000MHz, (Since RTX Voice forces to VRAM run at -200MHz, in order to compansate the deficit +200MHz OC applied for RTX Voice enabled tests on MSI Afterburner.)

DLSS OFF, RTX VOICE OFF, 47 FPS, 100.00%
DLSS OFF, RTX VOICE ON, 43 FPS, 91.4%
DLSS ON, RTX VOICE OFF, 83 FPS, 176.5%
DLSS ON, RTX VOICE ON, 76 FPS, 161.7%

I'm not sure if Nvidia Nsight should even capture RTX Voice foot print on GPU Trace but haven't seen any additional FP16+Tensor overhead. But there's relatively large difference on "Unused SM Warp Slots" metric 9.6% vs 14.8%

RTX Voice Tensor Load Comparison
 
Back
Top