GPUs Graphics Reviews

KFA2 GeForce RTX 3080 SG 10 GB in test – not quite as quiet, but really cool

Even if the GeForce RTX 3080 is generally only available in smaller quantities at the market, I want to offer nevertheless briefly from the market start of the KFA2 RTX 3080 SG still a suitable test of this Custom solution. In contrast to the rather narrow Founders Edition, KFA2 relies on the well-known 2.5-slot design and puts a card on the shelves that is supposed to offer a good compromise between performance and cost. However, the test will still have to prove how well this balancing act has been achieved.

First of all, a few key data about the GA102 chip used. The heart of this card is the GA102-200 graphics processor. It is based on Samsung’s 8nm process node designed specifically for NVIDIA and has a total of 28 billion transistors. It measures just under 628 mm², which makes it the second largest gaming GPU ever produced directly under the Turing TU102 GPU. The new shader cores of the Ampere architecture are 2.7x faster, the new RT cores 1.7x faster, while the new Tensor cores are up to 2.7x faster than their counterparts on Turing GPUs. The RT core of the 2nd. The 3rd generation also delivers dedicated hardware-accelerated ray tracing performance and offers twice the intersection of rays/triangles while providing RT graphics and arithmetic operations.

For the GeForce RTX 3080, NVIDIA has enabled a total of 68 SM units on its flagship, resulting in a total of 8704 CUDA cores. In addition to CUDA cores, NVIDIA’s GeForce RTX 3080 is also equipped with next-generation RT cores (ray tracing), Tensor cores, and brand new SM or streaming multiprocessor units. In terms of memory, the GeForce RTX 3080 is equipped with 10GB of GDDR6X memory that can deliver memory speeds of 19.0Gbps. Together with a 320 bit interface this will provide a cumulative bandwidth of 760 Gbps.

The KFA2 RTX 3080 SG 10 GB

The case of the card is the well-known mix of light metal, plastic and magnetic hardware, which is haptically ok, and visually not too thick. The design with the 5.5 cm installation depth and the 4 mm backplate makes this card a 2.5 slot design with all known advantages and disadvantages. You can love it or hate it, but it doesn’t scare off in the form of this graphics card, the power comes from fuel, as is well known, and it has to be dissipated as waste heat.

With its 1350 grams, the Founders Edition is almost as heavy or light as the 1352 gram Founders Edition! Coincidences are… The length of 31.5 cm is already slightly above average, whereas the rather moderate installation height of only 10.5 cm from the upper edge of the PCIe slot with installed card to the top of the cover is nothing to be afraid of in a case. The funny 12-pin Micro-Fit 3.0 on the top side was replaced by two normal 8-pin sockets, the cooler was also cut out at this point to match. Of course, there is also a little bit of RGB lighting on the upper side.

On the one hand, the back side shows two threaded inserts for the optional attachment to a suitable graphic card stand or mount, on the other hand there are still mysterious looking sockets, but they aren’t that mysterious at all. Here you could also connect a “1-Click-Booster”, an optional 8cm fan module that you could plug into the air outlet on the back. For whatever reason. Impressive are also the 5 tailpipes in the shape of the 5 (of 6) 8 mm heatpipes.

Like NVIDIA, KFA2 does without the USB Type C on the slot panel, probably also because the VR hype has cooled down considerably in the meantime. Interestingly, AMD will re-record this feature on Big-Navi, while it seems to have already finished here. The HDMI 2.1 connection should not be missing, and of course the three current DisplayPorts. The large cooling openings show that it could (and will) become quite warm.

The data in the BIOS is the same as in the Founders Edition, but with a power limit reduced to 320 watts. The 1440 MHz base and the 1710 MHz boost clock are already as well known as the 1188 MHz memory clock and the memory expansion with 10 GB at the 320-bit interphase. In direct comparison to the other new cards, the RTX 3080 with the 3090 is at the top end of the food chain and the performance-related distance of the 699 euro card to the 1499 euro expensive GeForce RTX 3090 is significantly smaller than the price difference. Here I also have a table for all statisticians among you, before it really gets going on the next page.

  GeForce RTX 3060 (Ti) GeForce RTX 3070 KFA2 RTX 3080 SG
GeForce RTX 3090
GPU GA104-200 GA104-300 GA102-200 GA102-300
Process node Samsung 8nm Samsung 8nm Samsung 8nm Samsung 8nm
The Size 395.2 mm2 395.2 mm2 628.4 mm2 628.4 mm2
Transistors 17.4 billion 17.4 billion 28 billion 28 billion
CUDA Cores 4864 5888 8704 10496
TMUs/ROPs TBA TBA 272 / 96 TBA
Tensor/RT 152 / 38 184 / 46 272 / 68 328 / 82
Basic Cycle
TBA 1500 MHz 1440 MHz 1400 MHz
Boost Clock
TBA 1730 MHz 1710 MHz 1700 MHz
FP32 Compute TBA 20 TFLOPs 30 TFLOPs 36 TFLOPs
RT TFLOPs TBA 40 TFLOPs 58 TFLOPs 69 TFLOPs
Tensor-TOPs TBA 163 TOPs 238 TOPs 285 TOPs
Memory 8 GB GDDR6 8 GDDR6 10 GB GDDR6X 24 GB GDDR6X
Interface 256-bit 256-bit 320-bit 384-bit
Throughput 14 Gbps 14 Gbps 19 Gbps 19.5 Gbps
Bandwidth 448 Gbps 448 Gbps 760 Gbps 936 Gbps
TGP 180W? 220W 320W 350W
Launch 10/2020 ? 15.10.2020 17.09.2020 24.09.2020

Test system and evaluation software

The benchmark system is new and is now no longer in the laboratory, but back in the editing room. I now also rely on PCIe 4.0, the matching X570 motherboard in the form of a MSI MEG X570 Godlike and a selected Ryzen 9 3900XT, which is water cooled and overclocked up to 4.5 GHz. In addition, there is the matching DDR4 3600 RAM from G.SKILL in the form of the TridentZ Neo, as well as several fast NVMe SSDs. For direct logging during all games and applications I use NVIDIA’s PCAD, which increases the comfort tremendously.

The measurement of power consumption and other things takes place here in the special laboratory on a redundant and identical in every detail test system then double-tracked using high-resolution oscillograph technology…

…and the self-created, MCU-based measurement setup for motherboards graphics cards (pictures below), where in the end the thermographic infrared images are also taken with a high-resolution industrial camera in an air-conditioned room. The audio measurements take place outside in my chamber.

I have also summarized the individual components of the test system in tabular form:

Test System and Equipment
Hardware:
AMD Ryzen 9 3900XT @4.5 GHz
MSI MEG X570 Godlike
2x 16 GB G.SKILL TridentZ Neo RGB DDR4 3600, CL
1x 2 TByte Aorus (NVMe System SSD, PCIe Gen. 4)
1x 500 GB Toshiba RC500
1x Seagate FastSSD Portable USB-C
Seasonic Prime 1300 Watt Titanium PSU
Cooling:
Alphacool Eisblock XPX Pro
Alphacool Eiswolf (modified)
Thermal Grizzly Kryonaut
Case:
Raijintek Paean
Monitor: BenQ PD3220U
Power Consumption:
Oscilloscope-based system:
Non-contact direct current measurement on PCIe slot (riser card)
Non-contact direct current measurement at the external PCIe power supply
Direct voltage measurement at the respective connectors and at the power supply unit
2x Rohde & Schwarz HMO 3054, 500 MHz multichannel oscilloscope with memory function
4x Rohde & Schwarz HZO50, current clamp adapter (1 mA to 30 A, 100 KHz, DC)
4x Rohde & Schwarz HZ355, probe (10:1, 500 MHz)
1x Rohde & Schwarz HMC 8012, HiRes digital multimeter with memory function

MCU-based shunt measuring (own build, Powenetics software)
Up to 10 channels (max. 100 values per second)
Special riser card with shunts for the PCIe x16 Slot (PEG)

NVIDIA PCAT and FrameView 1.1

Thermal Imager:
1x Optris PI640 + 2x Xi400 Thermal Imagers
Pix Connect Software
Type K Class 1 thermal sensors (up to 4 channels)
Acoustics:
NTI Audio M2211 (with calibration file)
Steinberg UR12 (with phantom power for the microphones)
Creative X7, Smaart v.7
Own anechoic chamber, 3.5 x 1.8 x 2.2 m (LxTxH)
Axial measurements, perpendicular to the centre of the sound source(s), measuring distance 50 cm
Noise emission in dBA (slow) as RTA measurement
Frequency spectrum as graphic
OS: Windows 10 Pro (2004, all updates, current certified drivers, NVIDIA 456.55)

Danke für die Spende



Du fandest, der Beitrag war interessant und möchtest uns unterstützen? Klasse!

Hier erfährst Du, wie: Hier spenden.

Hier kannst Du per PayPal spenden.

About the author

Igor Wallossek

Editor-in-chief and name-giver of igor'sLAB as the content successor of Tom's Hardware Germany, whose license was returned in June 2019 in order to better meet the qualitative demands of web content and challenges of new media such as YouTube with its own channel.

Computer nerd since 1983, audio freak since 1979 and pretty much open to anything with a plug or battery for over 50 years.

Follow Igor:
YouTube Facebook Instagram Twitter

Werbung

Werbung