Palit GeForce RTX 3090 GamingPRO OC 24G review
Palit has returned with their GamingPRO series graphics card, over the next 30 or so pages we review the Palit GeForce RTX 3090 GamingPRO OC armed with a muscular cooler and 24 GB of the fastest GDDR6X memory your money can get you. Well, that and a whopping 10469 Shader cores combined with Raytracing and Tensor cores, will it make a difference?
Armed with a shader core count that will make at least one your eyebrows frown with a nearly nauseating 24 GB of blazingly fast GDDR6X graphics memory. The GeForce RTX 3080 already is a smoking hot product, but of course, but NVIDIA is NVIDIA and decided to go stretch their legs a little more. One thing that needs to be stated, though, the product like shown today is all about Ultra HD and higher resolutions. It is an RTX Titan replacement or successor, I should say. I write all this before testing the product, but already understand that it will be hard to show where this product will make real sense. The good news is, for me at least, a product doesn’t have to really make sense in order to be appreciable. I mean you like Ferrari as well eh? (albeit right now that probably right now that is not the best example). Well, Tesla maybe .. a model S is out of range for many, but man they drive nice, accelerate fast and you get that feel you’re driving something from Star Trek. That said, in that realm, we feel the RTX 3090 will position itself, a niche.
It was 2017 when Ampere as a GPU architecture surfaced on the web and, up-to earlier this year, NVIDIA had not listed this name in any of its roadmaps on the consumer side. It was with military-level secrecy that the Ampere consumer part was developed. Ampere, of course, is the base unit of electric current in the international system of units. But the GPU is named after André-Marie Ampère, a French mathematician and physicist, considered the father of electrodynamics. NVIDIA has a track record of naming their GPU architectures after mathematicians and physicists or figures from closely related fields, to name a few; Pascal, Fermi, Kepler, Maxwell and, more recently, Turing. While it was no secret that the new GPUs would be based on Ampere, we’ve seen much discussion about fabrication nodes, architecture and specifications. Still, everybody seems to have forgotten that Ampere already launched earlier this year for the HPC market. The very first product based on Ampere was the NVIDIA Tesla A100, outfitted with a GA100 Ampere GPU based on 7nm fabricated at TSMC; that product holds 54 billion transistors and has 6912 Shader cores.
On September 1st of the year, 2020 NVIDIA announced three initial Ampere graphics cards in its first launch wave. A week before announcements, specifications of the GeForce RTX 3080 and 3090 took a twist; the shader core count mysteriously doubled up from what everybody expected. The GPUs are fabricated on an 8nm node derived from Samsung. This process is a further development of Samsung’s 10nm process, no EUV is applied in production just yet. The first wave of announcements would see the GeForce RTX 3080 and 3090 being released first and, as a bit of a surprise, the GeForce RTX 3070 would be arriving in roughly the same timeframe as well. The initial launch of Ampere for consumers entails the GeForce RTX 3070 8GB GDDR6, RTX 3080 10GB GDDR6X, and what we test today, the 24GB GDDR6X based premium flagship, the mighty mo, the GeForce RTX 3090. The lineup nearly doubles Raytracing performance with Gen2 ray-tracing cores and 3rd iteration Tensor cores. These cards will all be PCIe 4.0 interface compatible and offer HDMI 2.1 and DisplayPort 1.4a, but most importantly, is that exorbitant shader processor count (referred to as CUDA cores by NVIDIA), passing the 10K marker, nobody .. not even us saw that one coming.
The NVIDIA GA102 GPU is used initially for two products, the GeForce RTX 3080 and 3090 graphics cards. And it is one big GPU die and product overa;; alright, the 3090’s GA102-300-A1 GPU is armed with 10.496 Shader processors and 28 billion transistors. And no, that’s not even the fully unlocked product. FYI: the GeForce RTX 3080 is listed as having 8.704 Shader cores and the GeForce RTX 3070 (GA104) will bring 5.888 Shader cores to the table. In this review, we’ll check out the mother of them all, the GeForce RTX 3090, paired with 24GB of all that GDDR6X graphics memory, 24 GB, and 10K+ Shader cores.
Palit GeForce RTX 3090 GamingPRO OC 24G
The Palit GeForce RTX 3090 GamingPRO OC 24G is again fitted with that NVIDIA GA102 GPU, this time the revision 300 GPU SKU; it has an almost unimaginable 10.469 Shader cores activated and is paired with a staggering 24GB of all new GDDR6X graphics memory running at 19.5 Gbps, values that have been unprecedented until now. Palit applied a semi-passive design with three fans that start to spin and cool once the GPU warms up, and thus remains passive in idle load situations. The card is fitted with a duo of 8-pin power connectors. The card is rated with a 350W power draw, similar to the the reference (FE) design. This indicates it will perform fairly close to that FE product from NVIDIA. Weighing in at roughly 1500g this kidalso has been sized quite normally at 294 x 112 x 60 mm. The out of the box boost clock for this product is 1725 MHz (1695 MHz is the reference clock) but rated at 350 Watts it will perform like that FE as power distribution is the more dominant factor. Have a peek, and then let’s head onwards into the photoshoot. She’s quite pretty looking.