NVIDIA GeForce 705M

Power and Cooling — TDP, PSU recommendation, connector requirements

The NVIDIA GeForce 705M is a 15 W part, placing it firmly in the ultra-low-power segment of the mobile GPU landscape. This TDP figure is the single most defining characteristic for system integration, as it allows the chip to be deployed in thin-and-light laptops without active cooling solutions or substantial heatsinks. The data indicates the 705M draws so little power that it requires no external power connectors — the “None” entry in the power connector field confirms that all power is delivered through the PCIe 2.0 x16 bus interface itself. For a mobile implementation, this means the GPU is entirely dependent on the motherboard's power delivery design, and there is no separate PSU recommendation listed in the fact pack, which is typical for integrated-class graphics in laptops.

The 28 nm process node from TSMC, with 585 million transistors on a 116 mm² die, yields a transistor density of 5.0M per mm². This is a modest density by modern standards, but for 2013-era mobile hardware, it strikes a balance between leakage current and switching efficiency. The Fermi 2.0 architecture, despite being a generation old by the time of this chip's release, was known for its power inefficiency relative to newer designs — yet the 15 W TDP suggests NVIDIA aggressively clocked and voltage-limited this part to fit the mobile envelope. The pixel rate of 1.476 GPixel/s and texture rate of 5.904 GTexel/s are consistent with a chip that prioritizes battery life over raw throughput.

Given the absence of a suggested PSU in the fact pack, the 705M should be viewed as a component that never taxes a power supply. Even the weakest laptop adapters from its era would have ample headroom. The slot width is listed as “IGP” — integrated graphics platform — which reinforces that this GPU is soldered to the motherboard rather than being a discrete, replaceable module. The display outputs are “Portable Device Dependent,” meaning the actual connectors vary by laptop model, but the GPU itself does not dictate specific output configurations. For anyone analyzing this part for a retro build or upgrade path, the takeaway is simple: the 705M is thermally and electrically undemanding, but it is also not user-serviceable or upgradeable in any conventional sense.

Memory Subsystem — VRAM size/type, bus width, bandwidth and what it means for high resolutions

The GeForce 705M ships with 1024 MB of DDR3 memory, which was the entry-level capacity for mobile GPUs in its generation. The memory clock is 900 MHz, translating to 1800 Mbps effective due to DDR (double data rate) signaling. The bus width is a narrow 64 bit, which severely constrains memory throughput. Total bandwidth is 14.40 GB/s — a figure that was already modest at launch and is now borderline obsolete for any modern workload.

The 64-bit bus is the critical bottleneck here. With only 14.40 GB/s of bandwidth, the GPU cannot feed its 48 shading units efficiently at high resolutions. At 1080p or above, texture fetches and framebuffer writes will saturate the memory bus long before the compute units reach their limits. The 1024 MB VRAM capacity is also restrictive: modern games with high-resolution texture packs will exceed this allocation, forcing the driver to swap data between VRAM and system memory over the PCIe 2.0 x16 link, which further compounds the bandwidth problem. The fact that the memory type is DDR3 rather than GDDR5 means the chip uses lower-clocked, higher-latency memory modules, which is typical for power-constrained designs but worsens the performance ceiling.

For high-resolution gaming, the data paints a grim picture. Even at 720p, the 14.40 GB/s bandwidth will struggle with games that stream large textures. At 1080p, the pixel fill rate of 1.476 GPixel/s combined with the narrow bus means the 705M is effectively limited to very old titles, indie games, or 2D applications. The memory subsystem is the primary reason this GPU's benchmark percentile sits at 50 — it is not a failure of compute capability per se, but the memory design caps what the silicon can achieve.

Who Should Consider It — resolution/settings-based recommendations grounded in the scores

The GeForce 705M occupies a peculiar niche: it is an end-of-life product (production status confirms this) that was never intended for serious gaming. The benchmark data shows a 50th percentile ranking among all GPUs, which places it exactly at the median of a historical database — but that median includes integrated graphics and ancient parts. In practical terms, this GPU is suitable for basic desktop composition, video playback (though not high-bitrate 4K), and light productivity tasks.

At 720p resolution with low settings, the 705M can handle games from roughly 2010-2013 era, provided they are not GPU-intensive. Titles like older source-engine games, 2D platformers, or strategy games with modest graphics requirements might achieve playable frame rates. The FP32 compute of 141.7 GFLOPS is enough for basic physics and particle effects, but not for modern post-processing or high-detail geometry. At 1080p, the GPU is not recommended for any 3D gaming beyond the simplest titles — the memory bandwidth alone disqualifies it.

The target audience is not gamers. This is a GPU for office laptops, web browsing, and media consumption on small screens. The 1024 MB VRAM and 64-bit bus mean that even Windows desktop composition at high DPI could feel sluggish with multiple monitors. For anyone considering this chip today, the data suggests it is only viable for retro gaming at 720p or below, with all settings at minimum. The 15 W TDP makes it ideal for fanless designs, but the performance trade-off is severe.

How It Compares — position vs each nearest rival, one short paragraph per rival

The fact pack lists no nearest rivals for the GeForce 705M. The `nearestRivals` array is empty, and there are no benchmark scores or deltaPct values provided. This is unusual but not unprecedented for low-end mobile parts that were often OEM-only and never widely benchmarked. The absence of rival data means the 705M cannot be positioned against specific competing GPUs from AMD or Intel in this analysis.

Without rival comparisons, the only reference point is the global percentile of 50. This indicates that the 705M sits exactly in the middle of all GPUs ever recorded in the database — but given that the database includes decades of integrated and discrete parts, this median is heavily skewed toward older, weaker hardware. The 705M is likely outperformed by any modern integrated GPU (such as Intel's UHD series), but the fact pack does not provide data to confirm this. The lack of rivals also means no deltaPct values exist to quantify performance gaps. This makes the 705M a difficult part to recommend or even analyze comparatively, as there is no baseline to measure against.

Benchmark Performance — analyze scores vs rivals with exact % deltas

The `benchmarks` array in the fact pack is empty, and the `avgBenchmarkScore` is 0. This means there are no raw performance scores to analyze. The only quantitative performance metric is the percentile rank of 50, which is a relative measure rather than an absolute score. With no benchmark entries and no rival data, any percentage-based comparison is impossible.

What the data does provide is the theoretical peak rates: FP32 at 141.7 GFLOPS, pixel rate at 1.476 GPixel/s, and texture rate at 5.904 GTexel/s. These figures can be interpreted in isolation. A 141.7 GFLOPS FP32 throughput is roughly equivalent to a mid-range GPU from 2008-2010 era, but with a much lower power envelope. The pixel rate of 1.476 GPixel/s suggests that at 720p (1280x720 = 921,600 pixels), the GPU can theoretically fill approximately 1600 frames per second in a pure fill-rate test — but that ignores shader complexity and memory bandwidth. The texture rate of 5.904 GTexel/s is similarly theoretical; real-world games will not approach these limits due to the 14.40 GB/s memory bottleneck.

The absence of benchmark data is itself a finding. It suggests that the 705M was never subjected to standardized testing, or that its performance was so low as to be deemed unworthy of recording. The 0 avgBenchmarkScore is likely an artifact of no entries rather than a literal zero score. For analysts, this means the 705M must be evaluated on architectural merits alone, which are meager: 48 shading units, 8 TMUs, and 8 ROPs are the minimum configuration for a Fermi-based chip.

FAQ — 4-6 Q&A pairs, each answerable from FACT PACK data

Q: What is the TDP of the GeForce 705M and does it require external power?

A: The TDP is 15 W, and the power connectors field is listed as “None,” meaning all power comes from the PCIe 2.0 x16 bus.

Q: How much VRAM does the 705M have and what memory type does it use?

A: It has 1024 MB of DDR3 memory with a 64-bit bus, providing 14.40 GB/s of bandwidth at 1800 Mbps effective.

Q: What is the manufacturing process and transistor count?

A: The GPU is fabricated on a 28 nm TSMC process with 585 million transistors on a 116 mm² die, yielding a density of 5.0M transistors per mm².

Q: What API support does the 705M offer?

A: It supports DirectX 12 (11_0), OpenGL 4.6, and has no Vulkan support listed.

Q: What is the production status of this GPU?

A: The production status is “End-of-life,” with a release date of 2013-09-26 in the GeForce 700M generation.

Q: How many shading units, TMUs, and ROPs does the 705M have?

A: It has 48 shading units, 8 texture mapping units, and 8 render output units.

Ray Tracing and Feature Set — RT/tensor cores, API support from facts

The GeForce 705M has no ray tracing cores and no tensor cores — both fields are null in the fact pack. This places the GPU firmly in the pre-RTX era, where ray tracing was not a hardware feature but a software technique reserved for high-end workstation GPUs. The Fermi 2.0 architecture, dating back to 2010, has no dedicated hardware for either ray traversal or AI acceleration. Any ray tracing workload would be executed in software on the 48 shading units, which is impractical given the FP32 throughput of 141.7 GFLOPS.

The API support is the most forward-looking aspect of the 705M. It lists DirectX 12 (11_0), which means it can run DirectX 12 applications with feature level 11_0 — the baseline compatibility tier. This allows the GPU to execute DX12 titles, but without the advanced features like bindless resources or asynchronous compute that newer hardware supports. OpenGL 4.6 is also supported, which is surprisingly modern for a 2013 chip and ensures compatibility with most OpenGL-based applications. However, Vulkan support is null, meaning the GPU cannot use the modern cross-platform API that many current games rely on for efficient multi-threaded rendering.

The feature set lacks any Tensor or RT acceleration, so the 705M cannot participate in DLSS, ray-traced shadows, or any AI-enhanced rendering. For users seeking these features, the 705M is entirely unsuitable. The pixel rate of 1.476 GPixel/s and texture rate of 5.904 GTexel/s are the only other feature-related metrics, and they indicate that the GPU's fixed-function units are small but functional for basic 2D compositing and video decode. The architecture supports the DirectX feature level 11_0, which covers shader model 5.0, but nothing beyond that. In summary, the 705M is a legacy part with obsolete features, and its only modern concession is the DirectX 12 compatibility flag, which provides minimal practical benefit given the hardware's limitations.