NVIDIA GeForce RTX 3050 OEM Graphics Cards Feature Lower Cores Than The DIY Parts

First hi-res photos showing the GeForce RTX 3050's GA106-150 GPU Appear Online

NVIDIA's GeForce RTX 3050 graphics card is the most entry-level Ampere solution you can find in the current generation of the RTX 30 series lineup but it looks like the OEM version features cut-down specifications compared to the DIY variant.

NVIDIA GeForce RTX 3050 OEM Comes With 10% Lower Cores Compared To The DIY Graphics Card

The NVIDIA GeForce RTX 3050 graphics card comes in two variants, a GA106, and a GA107 model. Both models retain the same basic specifications such as 2506 CUDA cores, 80 TMUs, & 32 ROPs. Each variant features an 8 GB GDDR6 memory running across a 128-bit wide bus interface at 14 Gbps speeds for 224 GB/s bandwidth.

Related StoryHassan Mujtaba
NVIDIA & AMD GPU Prices Expected To Fall Further By End of August, Green Team To Issue Aggressive Cuts To Clear Inventory
The NVIDIA GeForce RTX 3050 OEM graphics card features fewer cores, lower clock speeds & features a standard OEM design versus the DIY variant. (Image Credits: ITHome).

Now, a new OEM variant of the NVIDIA GeForce RTX 3050 graphics card has appeared in China (via ITHome) which comes with a cut-down GA106 GPU core. This variant features 2304 Cores & a lower TMU / ROP count. The graphics card also rocks lower clock speeds of 1.51 GHz base (vs 1.55 GHz base) and a 1.76 GHz boost (versus 1.78 GHz boost) clock. The memory specification is kept intact & the card also rocks a single 8-pin connector to supplement its 130-ish watt TGP. Based on the cut-down specs, we can expect around 5-10% lower performance.

Users should be cautious of such OEM components with cut-down specifications as they can get lower performance for the product that has the same identical name as the full part. Currently, the NVIDIA GeForce RTX 3050 OEM only seems to be selling in Asian markets in OEM PC Builds but the pricing is said to be nearly the same as the DIY version which offers better specifications. The fact that NVIDIA and its OEM partners decided to keep the same naming scheme and not differentiate between the OEM and DIY parts is very misleading to consumers. This can be a tactic to put bad GA106 dies under the so-called 'GeForce RTX 3050'  and sell them to clueless buyers to get rid of additional inventory.

We advise you to check in with the OEM before purchasing any PC built with NVIDIA's GeForce RTX 3050 graphics card. The current price drop may lure customers in purchasing a slower graphics card than what they had hoped for.

NVIDIA GeForce RTX 30 Series Graphics Card Specifications

Graphics Card NameNVIDIA GeForce RTX 3090 TiNVIDIA GeForce RTX 3090NVIDIA GeForce RTX 3080 TiNVIDIA GeForce RTX 3080 12 GBNVIDIA GeForce RTX 3080NVIDIA GeForce RTX 3070 Ti 16 GBNVIDIA GeForce RTX 3070 TiNVIDIA GeForce RTX 3070NVIDIA GeForce RTX 3060 TiNVIDIA GeForce RTX 3060NVIDIA GeForce RTX 3050NVIDIA GeForce RTX 3050 OEM
GPU NameAmpere GA102-350?Ampere GA102-300Ampere GA102-225Ampere GA102-220?Ampere GA102-200Ampere GA104-400Ampere GA104-400Ampere GA104-300Ampere GA104-200
Ampere GA103-200
Ampere GA106-300Ampere GA106-150
Ampere GA107-300?
Ampere GA106-125?
Process NodeSamsung 8nmSamsung 8nmSamsung 8nmSamsung 8nmSamsung 8nmSamsung 8nmSamsung 8nmSamsung 8nmSamsung 8nmSamsung 8nmSamsung 8nmSamsung 8nm
Die Size628.4mm2628.4mm2628.4mm2628.4mm2628.4mm2395.2mm2395.2mm2395.2mm2395.2mm2 (GA104)276mm2276mm2 (GA106)276mm2 (GA106)
Transistors28 Billion28 Billion28 Billion28 Billion28 Billion17.4 Billion17.4 Billion17.4 Billion17.4 Billion (GA104)13.2 Billion13.2 Billion (GA106)13.2 Billion (GA106)
CUDA Cores107521049610240896087046144614458884864358425602304
TMUs / ROPs336 / 112328 / 112320 / 112280 / 104272 / 96184 / 96184 / 96184 / 96152 / 80112 / 6480 / 32TBD
Tensor / RT Cores336 / 84328 / 82320 / 80280 / 70272 / 68184 / 46184 / 46184 / 46152 / 38112 / 2880 / 20TBD
Base Clock1560 MHz1400 MHz1365 MHzTBA1440 MHzTBA1575 MHz1500 MHz1410 MHz1320 MHz1552 MHz1510 MHz
Boost Clock1860 MHz1700 MHz1665 MHzTBA1710 MHzTBA1770 MHz1730 MHz1665 MHz1780 MHz1777 MHz1760 MHz
FP32 Compute40 TFLOPs36 TFLOPs34 TFLOPsTBA30 TFLOPsTBA22 TFLOPs20 TFLOPs16 TFLOPs13 TFLOPs9.1 TFLOPs8.1 TFLOPs
RT TFLOPs74 RFLOPs69 TFLOPs67 TFLOPsTBA58 TFLOPsTBA44 TFLOPs40 TFLOPs32 TFLOPs25 TFLOPs18.2 TFLOPs16.2 TFLOPs
Tensor-TOPsTBA285 TOPs273 TOPsTBA238 TOPsTBA183 TOPs163 TOPs192 TOPs101 TOPs72.8 TOPs64.8 TOPs
Memory Capacity24 GB GDDR6X24 GB GDDR6X12 GB GDDR6X12 GB GDDR6X10 GB GDDR6X16 GB GDDR6X8 GB GDDR6X8 GB GDDR68 GB GDDR612 GB GDDR68 GB GDDR68 GB GDDR6
Memory Bus384-bit384-bit384-bit384-bit320-bit256-bit256-bit256-bit256-bit192-bit128-bit128-bit
Memory Speed21 Gbps19.5 Gbps19 Gbps19 Gbps19 Gbps21 Gbps19 Gbps14 Gbps14 Gbps16 Gbps14 Gbps14 Gbps
Bandwidth1008 GB/s936 GB/s912 Gbps912 Gbps760 GB/s672 GB/s608 GB/s448 GB/s448 GB/s384 GB/s224 GB/s224 GB/s
TGP450W350W350W350W320W~300W290W220W175W170W130W (GA106)
115W (GA107)
~130W
Price (MSRP / FE)TBD$1499 US$1199$999 US?$699 US$599 US?$599 US$499 US$399 US$329 US$249 USTBD
Launch (Availability)29th March 2022?24th September 20203rd June 202111th January 202217th September 2020Cancelled?10th June, 202129th October 20202nd December 202025th February 202127th January 20222022
WccfTech Tv
Subscribe
Filter videos by
Order