Jump to content

Alright all, I got I legit question. I hear everybody saying to wait until benchmarks results to buy a RTX 20 series gpu. The thing is. These new gpus have deep learning cores so won't that mean that every time you run a benchmark the score would get higher ? 

 

I might be stupid but in my opinion you can't trust those results any more because who can guaranty us that all the results they gonna show are from the first run.

 

Any thoughts on this?

Link to comment
https://linustechtips.com/topic/964470-rtx-20-series-benchmarks/
Share on other sites

Link to post
Share on other sites

Correct me if I'm wrong, but 'deep learning cores' means cores running algorithms Nvidia programmed into then through deep learning. In other words, 'deep learning' stuff done in the Nvidia company itself, not your graphics card.

CPU: i7-2600K 4751MHz 1.44V (software) --> 1.47V at the back of the socket Motherboard: Asrock Z77 Extreme4 (BCLK: 103.3MHz) CPU Cooler: Noctua NH-D15 RAM: Adata XPG 2x8GB DDR3 (XMP: 2133MHz 10-11-11-30 CR2, custom: 2203MHz 10-11-10-26 CR1 tRFC:230 tREFI:14000) GPU: Asus GTX 1070 Dual (Super Jetstream vbios, +70(2025-2088MHz)/+400(8.8Gbps)) SSD: Samsung 840 Pro 256GB (main boot drive), Transcend SSD370 128GB PSU: Seasonic X-660 80+ Gold Case: Antec P110 Silent, 5 intakes 1 exhaust Monitor: AOC G2460PF 1080p 144Hz (150Hz max w/ DP, 121Hz max w/ HDMI) TN panel Keyboard: Logitech G610 Orion (Cherry MX Blue) with SteelSeries Apex M260 keycaps Mouse: BenQ Zowie FK1

 

Model: HP Omen 17 17-an110ca CPU: i7-8750H (0.125V core & cache, 50mV SA undervolt) GPU: GTX 1060 6GB Mobile (+80/+450, 1650MHz~1750MHz 0.78V~0.85V) RAM: 8+8GB DDR4-2400 18-17-17-39 2T Storage: HP EX920 1TB PCIe x4 M.2 SSD + Crucial MX500 1TB 2.5" SATA SSD, 128GB Toshiba PCIe x2 M.2 SSD (KBG30ZMV128G) gone cooking externally, 1TB Seagate 7200RPM 2.5" HDD (ST1000LM049-2GH172) left outside Monitor: 1080p 126Hz IPS G-sync

 

Desktop benching:

Cinebench R15 Single thread:168 Multi-thread: 833 

SuperPi (v1.5 from Techpowerup, PI value output) 16K: 0.100s 1M: 8.255s 32M: 7m 45.93s

Link to post
Share on other sites

3 minutes ago, ShinchanNL said:

These new gpus have deep learning cores so won't that mean that every time you run a benchmark the score would get higher ? 

Thats not how it works. Those cores are only used for certain tasks and won't just make it faster. Things like games won't use the tensor cores at all.

 

Those cores basically do matrix multiplication and addition. They won't magically make it learn.

Link to post
Share on other sites

Deep Learning means its meant for scientific research and calculations relating to that.. not for improved performance with gaming each run... 

Community Standards

Please make sure to Quote me or @ me to see your reply!

Just because I am a Moderator does not mean I am always right. Please fact check me and verify my answer. 

 

"Beast Mode"

Ryzen 7 9800x3d | Arctic Liquid Freeze 3 Pro 360 | MSI X870 Tomahawk Wi-Fi | MSI RTX 5080 Gaming Trio OC | Gskill Flare X5 6000MT/s CL30

1tb WD Black SN850x NVMe | 4tb WD SN850x NVMe | Antec Flux Pro | Be Quiet Pure Power 13 M 1000w | OWC 10gb NIC

 

Dedicated Streaming Rig

 Ryzen 7 3700x | Asus B450-F Strix | 32gb Gskill Flare X 3200mhz | Corsair RM550x PSU | MSI Ventus 3060 12gb | 250gb 860 Evo m.2

Phanteks P300A |  Elgato HD60 Pro | Avermedia Live Gamer Duo | Avermedia 4k GC573 Capture Card

 

Link to post
Share on other sites

1 minute ago, Jurrunio said:

Correct me if I'm wrong, but 'deep learning cores' means cores running algorithms Nvidia programmed into then through deep learning. In other words, 'deep learning' stuff done in the Nvidia company itself, not your graphics card.

These cores are desiged to do deep learing on the gpu for programs like tensor flow. This is normally used to power large datacenters. The deep leariing code is done by the customer.

Link to post
Share on other sites

6 minutes ago, ShinchanNL said:

Alright all, I got I legit question. I hear everybody saying to wait until benchmarks results to buy a RTX 20 series gpu. The thing is. These new gpus have deep learning cores so won't that mean that every time you run a benchmark the score would get higher ? 

 

I might be stupid but in my opinion you can't trust those results any more because who can guaranty us that all the results they gonna show are from the first run.

 

Any thoughts on this?

it can be only used by software with deeplearning and ai support but usually benchmarks dont.

Link to post
Share on other sites

Sadly, no. Those cores are only meant for certain application and wont accelerate your gameplay. The reasok we say "wait for benchmarks" is that the new cards havent gotten any level performance benchmarks against older titles without raytracing, something Nvidia just doesnt do. Not to mention a whole month untill launch(and pre-orders during that time) when the press knows practically nothing of the performance of the cards. 

 

The new cores in the cards is very nice, but in many ways its an early adopter card. Which is great for advancing the industry, but doesnt offer much to the consumer

Link to post
Share on other sites

Just now, GoldenLag said:

Sadly, no. Those cores are only meant for certain application and wont accelerate your gameplay. The reasok we say "wait for benchmarks" is that the new cards havent gotten any level performance benchmarks against older titles without raytracing, something Nvidia just doesnt do. Not to mention a whole month untill launch(and pre-orders during that time) when the press knows practically nothing of the performance of the cards. 

 

The new cores in the cards is very nice, but in many ways its an early adopter card. Which is great for advancing the industry, but doesnt offer much to the consumer

Yeah. Guess I'll sit this generation out on my 1080.

Link to post
Share on other sites

1 minute ago, ShinchanNL said:

Yeah. Guess I'll sit this generation out on my 1080.

The performance jump from the 1080 is not worth the pricetag. You are better of waiting untill next gen when raytracing will have better implemantation not to mention performance. (The demoes seem to have had a framerate of about 30-60 fps, its early performance metrics though not expected to improve much)

Link to post
Share on other sites

I'm guessing you're thinking of DLSS (Deep learning Super Sampling)?


As far as I know, (and I've yet to read any really detailed info on DLSS yet, so I may be completely wrong) it's just a way to off-load super sampling/anti aliasing tasks from the cuda cores to the tensor cores. Take the load of AA tasks off of the cuda cores and it should free up a lot of processing power, and in turn the card can push higher FPS.  Similar to how you get better FPS if you lower or switch off Anti-Aliasing.

Quote

What’s interesting about DLSS is that it uses Turing’s Tensor cores to enable a deep learning-powered AI technique to dramatically improve image quality and remove jagged edge artifacts in games, with a minimal performance hit. Traditional anti-aliasing, or AA, can lean heavily on GPU processing resources and memory bandwidth, and may consume large amounts of frame buffer memory as well. Anti-aliasing often requires a frame to be rendered and re-rendered multiple times in order to smooth out edges of objects in a scene and the end result can be a costly performance hit. NVIDIA notes that its new DLSS anti-aliasing technique, however, uses machine learning of scenes and images (called inferencing) to improve image quality without the large performance hit associated with some traditional AA methods.

https://hothardware.com/news/nvidia-geforce-rtx-2080-performance-and-dlss


To answer your question, I'm sure that when the cards are released the reviewers will be testing and benchmarking performance with and without DLSS enabled. Also, DLSS has to be enabled by the game/benchmark that is running, so if the benchmark doesn't support it, then the card won't be able to take advantage of DLSS.

CPU: Intel i7 6700k  | Motherboard: Gigabyte Z170x Gaming 5 | RAM: 2x16GB 3000MHz Corsair Vengeance LPX | GPU: Gigabyte Aorus GTX 1080ti | PSU: Corsair RM750x (2018) | Case: BeQuiet SilentBase 800 | Cooler: Arctic Freezer 34 eSports | SSD: Samsung 970 Evo 500GB + Samsung 840 500GB + Crucial MX500 2TB | Monitor: Acer Predator XB271HU + Samsung BX2450

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×