GIGABYTE today introduced its GeForce RTX 3090 Eagle OC graphics card alongside NVIDIA's debut of the monstrous RTX 3090 "Ampere" GPU. The card combines NVIDIA's de facto flagship product based on Ampere, the RTX 3090, with a mammoth 24 GB of GDDR6X memory, factory-overclocked speeds, and a board design more for gamers who want to load their battlestations with the most powerful graphics option in the market. GIGABYTE introduced the Eagle brand of custom-design graphics cards covering the value-end of its product stack, which runs from here through to the WindForce OC series and eventually onto the coveted AORUS Gaming series geared toward enthusiasts and overclockers. That said, the Eagle shares many underpinnings with the WindForce OC series.
NVIDIA has restructured the top end of its product stack with the RTX 30-series. The $700 GeForce RTX 3080 launched last week is being referred to as a "flagship product," and it's shown beating not just its logical predecessor, the RTX 2080, but also the previous-generation flagship, the RTX 2080 Ti, while being a fair bit cheaper. The RTX 3090, on the other hand, is being labeled a "halo" product, a title previously held by the $2,500 TITAN RTX. With the RTX 3080 already offering 4K UHD gaming with RTX-on, the RTX 3090 transcends product-segments between gaming and professional visualization, with NVIDIA extensively talking to us about its benefits to creators, especially when combined with GeForce Studio drivers. The TITAN RTX itself was such an inter-stack product. So from this perspective, the RTX 3090 is exactly 50% pricier than the RTX 2080 Ti at launch, but also $1,000 less than the TITAN RTX.
As part of the product stack restructuring, NVIDIA leveraged a common "GA102" silicon for the RTX 3090 and RTX 3080, with the RTX 3080 heavily cut down from it, while the RTX 3090 nearly maxes it out. Last time around, the RTX 2080 and its refresh, the RTX 2080 Super, were based on the second-largest TU104 chip, while the RTX 2080 Ti is based on the larger TU102. Between the RTX 3080 and RTX 3090, NVIDIA carved itself large amounts of room for future segmentation.
NVIDIA created the GeForce RTX 3090 out of the "GA102" Ampere silicon by enabling all but one of its TPCs, nearly maxing it out with a staggering 10,496 CUDA cores, 328 tensor cores, 82 RT cores, 328 TMUs, and 112 ROPs. NVIDIA maxed out the chip's 384-bit wide memory interface to cram in 24 GB of it. Ampere represents the 2nd generation of NVIDIA's path-breaking RTX architecture that introduces real-time raytracing to the consumer segment by combining conventional raster 3D graphics with real-time raytraced elements, such as lighting, shadows, reflections, ambient-occlusion, and global illumination. The 2nd generation even introduces raytraced motion-blur and has fixed-function hardware just to pull this otherwise difficult effect off. Find more details about the architecture in our NVIDIA Ampere Architecture article.
Unlike TITAN-branded graphics cards, the RTX 3090 is open for NVIDIA's partners to come up with custom designs. GIGABYTE paired the RTX 3090 with a meaty dual fin-stack heatsink that uses six copper heat pipes, a copper base-plate, and three fans. Keeping up with the hot new design trend of this generation, the cooler is longer than the PCB, which means much of the airflow from one of the fans goes straight through the card and out of the backplate, mimicking the Dual-Axial Flow Through cooling solution of the NVIDIA Founders Edition cards. The Eagle OC comes with factory-overclocked GPU Boost speeds of 1725 MHz, while the memory is left untouched at 19.5 Gbps (GDDR6X effective). GIGABYTE is pricing the card at the $1499 MSRP.
GeForce RTX 3090 Market Segment Analysis
Price
Shader Units
ROPs
Core Clock
Boost Clock
Memory Clock
GPU
Transistors
Memory
GTX 1080 Ti
$650
3584
88
1481 MHz
1582 MHz
1376 MHz
GP102
12000M
11 GB, GDDR5X, 352-bit
RX 5700 XT
$370
2560
64
1605 MHz
1755 MHz
1750 MHz
Navi 10
10300M
8 GB, GDDR6, 256-bit
RTX 2070
$340
2304
64
1410 MHz
1620 MHz
1750 MHz
TU106
10800M
8 GB, GDDR6, 256-bit
RTX 2070 Super
$450
2560
64
1605 MHz
1770 MHz
1750 MHz
TU104
13600M
8 GB, GDDR6, 256-bit
Radeon VII
$680
3840
64
1802 MHz
N/A
1000 MHz
Vega 20
13230M
16 GB, HBM2, 4096-bit
RTX 2080
$600
2944
64
1515 MHz
1710 MHz
1750 MHz
TU104
13600M
8 GB, GDDR6, 256-bit
RTX 2080 Super
$690
3072
64
1650 MHz
1815 MHz
1940 MHz
TU104
13600M
8 GB, GDDR6, 256-bit
RTX 2080 Ti
$1000
4352
88
1350 MHz
1545 MHz
1750 MHz
TU102
18600M
11 GB, GDDR6, 352-bit
RTX 3070
$500
5888
96
1500 MHz
1725 MHz
1750 MHz
GA104
17400M
8 GB, GDDR6, 256-bit
RTX 3080
$700
8704
96
1440 MHz
1710 MHz
1188 MHz
GA102
28000M
10 GB, GDDR6X, 320-bit
RTX 3090
$1500
10496
112
1395 MHz
1695 MHz
1219 MHz
GA102
28000M
24 GB, GDDR6X, 384-bit
Gigabyte RTX 3090 Eagle OC
$1500
10496
112
1395 MHz
1725 MHz
1219 MHz
GA102
28000M
24 GB, GDDR6X, 384-bit
Packaging
The Card
The RTX 3090 Eagle OC has many design underpinnings from the company's WindForce OC line of graphics cards, but comes with a different cooler shroud design. Its design involves an aluminium dual fin-stack heatsink that's ventilated by three fans; two of these are 90 mm and one 80 mm. The one near the display connectors is the smaller of the three. The PCB is shorter than the cooler, so some of the airflow from the third fan flows through the backplate.
Dimensions of the card are 33 x 13 cm.
Installation requires three slots in your system.
Display connectivity options include three standard DisplayPort 1.4a and two HDMI 2.1. Interestingly, the USB-C port for VR headsets, which NVIDIA introduced on Turing Founders Editions, has been removed—guess it didn't take off as planned. The DisplayPort 1.4a outputs support Display Stream Compression (DSC) 1.2a, which lets you connect 4K displays at 120 Hz and 8K displays at 60 Hz. Ampere can drive two 8K displays at 60 Hz with just one cable per display.
Ampere is the first GPU to support HDMI 2.1, which increases bandwidth to 48 Gbps to support higher resolutions, like 4K144 and 8K30, with a single cable. With DSC, this goes up to 4K240 and 8K120. NVIDIA's new NVENC/NVDEC video engine is optimized to handle video tasks with minimal CPU load. The highlight here is added support for AV1 decode. Just like on Turing, you may also decode MPEG-2, VC1, VP8, VP9, H.264, and H.265 natively, at up to 8K@12-bit.
The encoder is identical to Turing. It supports H.264, H.265, and lossless at up to 8K@10-bit.
Unlike the NVIDIA Founders Edition card that introduces the new 12-pin power input, Gigabyte uses a pair of 8-pin PCIe power inputs. Combined with the PCIe slot power, this configuration is rated for 375 W.
The GeForce RTX 3090 supports SLI and features a newer-generation NVLink bridge interface, which means you can't use your NVLink bridge from your Turing cards. Be warned that with Ampere, NVIDIA isn't supporting SLI as in implicit multi-GPU (SLI as you know it), but explicit multi-GPU developed and supported by game developers instead. With multi-GPU game support being pretty much non-existent, this basically means SLI is dead. Perhaps creative and 3D modeling applications that support explicit multi-GPU can benefit from SLI.
Teardown
Disassembling the Gigabyte RTX 3090 Eagle OC is fairly straightforward. You undo a set of screws at the backplate and another set underneath, which has the cooler come off clean.
As we mentioned, the cooler is longer than the PCB, so rather than placing its two 8-pin PCIe power connectors where the PCB ends, which is roughly at two-thirds the length of the card, Gigabyte used an elaborate contraption that extends the 8-pin connectors to the edge of the card by implementing internal wiring and a proprietary 8-pin single-row connector internally. It looks to have been integrated with the backplate, but comes off when you undo a pair of screws—there's some extra-high gauge wiring helping things. This also opens up the possibility for Gigabyte to switch the power inputs to three 8-pins, or even the NVIDIA 12-pin, without changing the PCB. Even when integrated with the backplate, this contraption shouldn't get in the way of full-coverage water blocks designed for the card.
Gigabyte is using a metal backplate with the RTX 3090 Eagle OC, which pulls heat from the memory chips on the reverse side of the PCB using thermal pads.
High-resolution PCB Pictures
These pictures are for the convenience of volt modders and people who would like to see all the finer details on the PCB. Feel free to link back to us and use these in your articles or forum posts.
High-res versions are also available (front, back).