After leaks upon leaks spanning across two months, a new report every week, and multiple last-minute delays, NVIDIA has finally launched its new entry-level GPU, the GeForce GTX 1630. Previously, we covered the possible release date of this card that was leaked on the Chinese Board Channels. It looked legit and everything was in place with no delays in sight, and today, that leak materialized into reality.
NVIDIA has unveiled the latest addition to its GeForce family with a new card in the 16-series lineup. The GTX 1630 is intended to replace the legendary GTX 1050 Ti to become the default entry-level offering from the Green Team. As the name suggests, the card is very similar in nature to the GTX 1650, which in many ways is the bigger brother.
For starters, both the GTX 1630 and GTX 1650 use the same TU117 GPU, however the former is using a cut-down version of it (TU117-150 vs. TU117-300). That slices the CUDA Core count from 896 down to just 512. Same goes for the memory bus interface whose width has been cut down in half, from 128-bit to just 64-bit.
Even though, the memory capacity remains the same at 4GB of GDDR6 along with 12Gbps speed, the reduced bus interface width significantly worsens the card’s memory bandwidth. Compared to the GTX 1650’s 192GB/s bandwidth, the GTX 1630 only has 96GB/s, which will effect performance severely, especially in a GPU this lower-end.
Memory bandwidth is more crucial than ever in games due to higher-quality assets needing to be loaded in faster. Therefore, having a less-than-100GB/s bandwidth really makes this GPU a tough recommendation for gaming in 2022. Heck, due to its lesser CUDA Cores and memory situation, the GTX 1630 is actually slower than the GTX 1050 Ti, the card it literally intends to replace.
The similarities between the GTX 1650 and GTX 1630 grow even deeper when you learn that manufacturers and AIBs have rehashed GTX 1650 designs to power their new GTX 1630 variants. The same PCB, same cooling system, same video outputs, all of it.
Moreover, just like the 1650, the GTX 1630 also features a 75W TGP, which means it requires no external power connectors. Though, some aftermarket variants with beefier cooling solutions and factory overclocks come with higher power limits, requiring them to equip the GPU with a 6-pin power connector.
Third-party variants and benchmarks
Speaking of aftermarket variants, so far EVGA, ZOTAC, GIGABYTE, PALIT, INNO3D, Gainward, MSI, PNY, and COLORFUL have announced their GTX 1630 models. We’ve already covered COLORFUL’s BattleAx design in detail in a previous article.
Whereas, another variant, the Gainward GTX 1630 Ghost has been tested by TechPowerUP already and it’s the only official review of the card out right now. To summarize, the GTX 1630 is about 57% slower than AMD‘s RX 6400 while costing almost the same, and consuming very similar amounts of power. Take a look at these benchmark results:
In overall gaming performance, the GTX 1630 only managed to outshine some really, really old GPUs that weren’t well-reviewed for even their time. The only metric where the GTX 1630 stands toe-to-toe against the RX 6400 is power consumption in gaming which is actually 2W lower on the NVIDIA offering. But overall, the card is still less efficient than the RX 6400.
These benchmarks did not include the GTX 1050 Ti which is a bummer, as that’s the GPU to beat for the GTX 1630 for it to be able to justify its existence. Regardless, on-paper specs tell us it’s worse but maybe generational improvements made with Turing over Pascal can somehow save it. For now, though, this is the best we have.
Pricing and availability
As for the pricing, the card doesn’t have an official MSRP in USD as it’s mainly aimed at Asia Pacific markets. The model will, however, retail for 1000-1100 RMB in China, which roughly comes out to around $150. COLORFUL and EVGA has listed their custom variants for sale at $169 and a whooping $199, respectively.
EVGA’s GTX 1630 is currently more expensive than their GTX 1650 model, which is just absurd. So, don’t be surprised if this card sells for around $200 in the States.
All in all, the GTX 1630 is certainly an interestingly release. It’s such a low-powered GPU that most people won’t even come across its announcement (or existence) but it is expected to populate the OEM market very quickly. System Integrators often put entry-level GPUs inside their PCs, just for the sake of providing a video output and for that, this card is more than perfectly fine.
NVIDIA’s main goal is to basically phase out Turing as fast as possible. They have an abundance of Turning chips lying around which clog up the production channels for current-gen and next-gen GPUs, so getting rid of it with new launches is the only choice. As for the actual quality of the product, well, it’s no RTX 3090, let me tell you that.