AMD issued a preemptive response to nVidia's new GTX 1050 and GTX 1050 Ti, and they did it by dropping the RX 460 MSRP to $100 and RX 470 MSRP to $170. The price reduction's issuance is to battle the GTX 1050, a $110 MSRP card, and GTX 1050 Ti, a $140-$170 card. These new Pascal-family devices are targeted most appropriately at the 1080p crowd, where the GTX 1060 and up were all capable performers for most 1440p gaming scenarios. AMD has held the sub-$200 market since the launch of its RX 480 4GB, RX 470, and RX 460 through the summer months, and is just now seeing its competition's gaze shift from the high-end.
Today, we've got thermal, power, and overclocking benchmarks for the GTX 1050 and GTX 1050 Ti cards. Our FPS benchmarks look at the GTX 1050 OC and GTX 1050 Ti Gaming X cards versus the RX 460, RX 470, GTX 950, 750 Ti, and 1060 devices. Some of our charts include higher-end devices as well, though you'd be better off looking at our GTX 1060 or RX 480 content for more on that. Here's a list of recent and relevant articles:
MSI GTX 1050 OC & GTX 1050 Ti Gaming X Review [Video]
Continue on for the full review article, as always. The video is above.
GTX 1050 Ti vs. GTX 1050 Specs, & 1060, 1070, & 1080
NVIDIA Pascal Specs Comparison | ||||||
GTX 1050 | GTX 1050 Ti | GTX 1060 3GB | GTX 1060 6GB | GTX 1070 | GTX 1080 | |
GPU | GP107-300 | GP107-400 | GP106 Pascal | GP106 Pascal | GP104-200 Pascal | GP104-400 Pascal |
Transistor Count | 3.3B | 3.3B | 4.4B | 4.4B | 7.2B | 7.2B |
Fab Process | 14nm | 14nm | 16nm FinFET | 16nm FinFET | 16nm FinFET | 16nm FinFET |
CUDA Cores | 640 | 768 | 1152 | 1280 | 1920 | 2560 |
GPCs | 2 | 2 | 2 | 2 | 3 | 4 |
SMs | 5 | 6 | 9 | 10 | 15 | 20 |
TPCs | 5 | 6 | 9 | 10 | 15 | 20 |
TMUs | 40 | 48 | 72 | 80 | 120 | 160 |
ROPs | 32 | 32 | 48 (?) | 48 | 64 | 64 |
Core Clock | 1354MHz | 1290MHz | 1506MHz | 1506MHz | 1506MHz | 1607MHz |
Boost Clock | 1455MHz | 1392MHz | 1708MHz | 1708MHz | 1683MHz | 1733MHz |
FP32 TFLOPs | 1.86TFLOPs | 2.14TFLOPs | 3.85TFLOPs | 4.4FLOPs | 6.5TFLOPs | 9TFLOPs |
Memory Type | GDDR5 | GDDR5 | GDDR5 | GDDR5 | GDDR5 | GDDR5X |
Memory Capacity | 2GB | 4GB | 3GB | 6GB | 8GB | 8GB |
Memory Clock | 7Gbps | 7Gbps | 8Gbps | 8Gbps | 8Gbps | 10Gbps GDDR5X |
Memory Interface | 128-bit | 128-bit | 192-bit | 192-bit | 256-bit | 256-bit |
Memory Bandwidth | 112GB/s | 112GB/s | 192GB/s | 192GB/s | 256GB/s | 320.32GB/s |
TDP | 75W | 75W | 120W | 120W | 150W | 180W |
Power Connectors | None | None | 1x 6-pin | 1x 6-pin | 1x 8-pin | 1x 8-pin |
Release Date | November, 2016 | 10/25/2016 | August, 2016 | 7/19/2016 | 6/10/2016 | 5/27/2016 |
Release Price | $110 | $140 | MSRP: $200 | Reference: $300 MSRP: $250 |
Reference: $450 MSRP: $380 |
Reference: $700 MSRP: $600 |
GTX 1050 & GTX 1050 Ti GP107-400 & GP107-300 Architecture
The GTX 1050 Ti we're working with is an MSI 1050 Ti Gaming X, with a price-point a bit higher than the $140 MSRP of the GTX 1050 Ti. The Gaming X should land at $165 – or about where the RX 470 is supposed to be – and the GTX 1050 base model we're using is priced at $110. The GTX 1050 Ti begins availability today, October 25, and the GTX 1050 non-Ti will become available “on or before November 8.” MSRP is $110 for the GTX 1050, $140 for the GTX 1050 Ti, and AIB partners will offer cards priced above MSRP to fill various markets.
Recapping the architecture before getting further, the GTX 1050 and GTX 1050 Ti cards are both operating on a small, 132mm^2 die size. This is among the smallest of the current-gen GPUs we've looked at to-date. Thermal dissipation potential across such a small surface is reduced, but the GPU is also a low-power chip and requires only 75W of cooling power. Most GTX 1050 and GTX 1050 Ti units will draw their power through the PCI-e bus, with a few overclocking models (like the Gaming X that we received) working with an additional 6-pin input for more juice availability.
The GP107 GPU comes in two flavors, just like the GP104 GPU used in the GTX 1080 and GTX 1070. GP107-400 is the GTX 1050 Ti's GPU, while GP107-300 is the GTX 1050's GPU. The two GPUs are separated by 1SM (6 simultaneous multiprocessors on the GTX 1050 Ti, 5 on the GTX 1050), naturally indicating a difference of 128 CUDA cores and 8 TMUs. The graphics card SKUs are further differentiated by VRAM capacity, with the GTX 1050 at 2GB (and only 2GB), and the GTX 1050 Ti at 4GB (and only 4GB).
Pascal's architecture remains the same on GP107 as it did with GP106, GP104, and GP102, though is different from the GP100 Pascal unit that uses different SM layout and has greater focus on FP16 & FP64 CUDA cores. One noteworthy point is that nVidia is building the GTX 1050 and 1050 Ti cards on 14nm fabrication process rather than the 16nm TSMC process most heavily distributed in the higher-end devices, though this is less significant than it sounds. NVidia switches supply based on availability to ensure the chain can keep up with demand. This is also why some GTX 1070s shipped with Samsung memory initially, but made a switch-over to Micron memory.
GP107 is a small chip. It's running two GPCs, just like the GTX 1060. The memory subsystem is built on a 128-bit wide bus and uses memory with speeds minimally at 7Gbps. This is true for both the GTX 1050 and 1050 Ti, each of which has a base memory bandwidth of 112GB/s.
Clock-rates are advertised as overclocking upwards of 1900MHz, something we'll test later, with the boosted core clock-rate at 1354MHz for the GTX 1050 and 1392MHz for the GTX 1050 Ti. GP107 uses the same clock-gating, heuristics, and memory compression techniques for per-bit power savings as the rest of Pascal. We more heavily defined Pascal architecture in our GTX 1080 (here) and GTX 1070 reviews, if interested to learn more.
Continue to Page 2 for testing methodology.