AMD Radeon R7 350 OEM

The AMD Radeon R7 350 OEM is a 28 nm GCN 1.0 part built on the Oland chip, released on 2015-05-04, and now listed as end-of-life. It belongs to the Pirate Islands (R7 300) generation and uses a 77 mm² die with 950 million transistors, a density of 12.3M / mm². The fact pack contains no benchmark entries, an average benchmark score of 0, and an empty nearestRivals array, leaving the 50th percentile among all GPUs as the only global comparison point.

Benchmark Performance

The data set provides no individual benchmark scores and no nearestRivals entries, so exact percentage deltas against named products cannot be computed from this record. The avgBenchmarkScore field is 0, which is the value associated with the empty benchmark list rather than a tested score. The percentileVsAllGpus field is 50, placing the R7 350 OEM at the midpoint of the GPU distribution in the database. That midpoint is a relative position, but it is not supported by measured runs in the data.

What can be quantified from the fact pack is the specification-derived throughput. The pixel rate is 8.400 GPixel/s, the texture rate is 25.20 GTexel/s, and FP32 compute is 806.4 GFLOPS. These are the hardware’s rated limits. The base clock is 1000 MHz and the boost clock is 1050 MHz. The listed pixel rate aligns with 8 ROPs at 1050 MHz, the texture rate aligns with 24 TMUs at 1050 MHz, and the FP32 rate aligns with 384 shading units at 1050 MHz. That makes the boost clock the reference point for the throughput fields in the fact pack.

With 384 shading units, 24 texture units, and 8 ROPs, the hardware is oriented toward shading and texture work rather than heavy fillrate. Benchmarks would be needed to translate that orientation into real application performance, but the data does not include any. The 50th percentile position in the broader database, taken alongside the empty benchmark field, implies a middle-of-the-pack database placement. Since no nearestRivals are listed, no exact percentage deltas can be stated. In this specific record, the percentile is the only comparative score, and it is not accompanied by a benchmark distribution. The performance story is incomplete rather than negative.

Who Should Consider It

The memory and render configuration define the target use case. The R7 350 OEM has 2 GB of GDDR5 memory, a 128-bit bus, and 73.60 GB/s of bandwidth. On the render side, it has 8 ROPs and 384 shading units. This is a low-throughput profile best suited to modest rendering workloads. Users who keep texture quality and resolution within the 2 GB frame buffer are more likely to stay within the card’s envelope than users who push high-resolution assets.

The absence of benchmark scores prevents a hard resolution and settings recommendation. The data does show a 65 W TDP and no power connectors, which points to a card for systems with limited power supply headroom. The suggested PSU is 250 W, so the data positions it inside a low-wattage system context. It is an OEM-style card, based on the name and the display output suite. For users with light rendering needs, the 2 GB capacity and 73.60 GB/s bandwidth are the relevant limits. For users with heavier expectations, the empty benchmark record provides no measured confirmation.

Ray Tracing and Feature Set

The fact pack lists rtCores as null and tensorCores as null. That means no dedicated ray tracing hardware and no tensor processing cores are recorded for this GPU. The architecture is GCN 1.0. The API list includes DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. These APIs provide modern framework support, but the null RT and tensor fields mean no dedicated acceleration blocks are in the data.

The feature set is therefore a split picture: modern API compatibility alongside a GCN 1.0 core without RT or tensor hardware. The API list alone does not establish ray tracing capability because the rtCores field is null. Vulkan 1.2.170 is listed, and OpenGL 4.6 is present. The display outputs are 1x DVI, 1x HDMI 1.4a, and 1x VGA.

Power and Cooling

The power profile of the R7 350 OEM is low. The TDP is 65 W, and the suggested PSU is 250 W. The power connector field is None, so the card does not need external power cables; it draws through the PCIe slot. The board is dual-slot, as recorded in the slotWidth field. The bus interface is PCIe 3.0 x8, as recorded in the busInterface field.

For power and cooling, this is a simple installation scenario. A 250 W PSU is a modest requirement, and the lack of a power connector removes cable management concerns. The dual-slot cooler is the main physical space consideration. The data does not include length, height, or width dimensions, so physical clearance cannot be assessed from this record. The low TDP of 65 W suggests a modest cooling requirement.

FAQ

Q: Does the AMD Radeon R7 350 OEM support hardware ray tracing?

A: No. The fact pack lists rtCores and tensorCores as null, so no dedicated ray tracing or tensor hardware is recorded.

Q: What are the memory specifications?

A: It uses 2 GB of GDDR5 on a 128-bit bus, with a 1150 MHz memory clock and 4.6 Gbps effective data rate, providing 73.60 GB/s of bandwidth.

Q: What power supply is suggested?

A: The suggested PSU is 250 W, and the TDP is 65 W. The power connector field is None, so no auxiliary power connector is needed.

Q: What display outputs does the card have?

A: The display outputs are 1x DVI, 1x HDMI 1.4a, and 1x VGA.

Q: What graphics APIs are supported?

A: The card supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170.

Q: Is the card still in production?

A: No. Its production status is end-of-life, and its release date is 2015-05-04.

Memory Subsystem

The memory subsystem is a central part of the R7 350 OEM’s identity. It uses 2 GB of GDDR5 on a 128-bit bus. The memory clock is 1150 MHz, and the effective data rate is 4.6 Gbps. The resulting peak bandwidth is 73.60 GB/s. This is a modest memory pipe within the context of the card’s own specifications.

At high resolutions, the 2 GB capacity is a hard cap on how much frame data can be held on the GPU. The 128-bit bus width and 73.60 GB/s bandwidth set the speed of that memory pipe. The effective 4.6 Gbps data rate is the advertised transfer rate; the bus width translates that into total bandwidth. For lighter detail settings and lower resolution workloads, the 2 GB and 73.60 GB/s combination can be adequate. For high-resolution assets, the capacity and bandwidth are limits. The memory subsystem is small, narrow, and matched to a low-power renderer.

How It Compares

The nearestRivals array in this fact pack is empty, so there are no named rivals to write about. The only comparative figure is percentileVsAllGpus, which is 50. This is the midpoint of the all-GPU distribution in the database. The average benchmark score of 0 reinforces that no measured rival deltas are available.

In the product lineage, the R7 350 OEM sits between Volcanic Islands as its predecessor and Arctic Islands as its successor. That is a generational placement, not a performance score. Without nearestRivals, any claim that it beats a specific rival cannot be sourced from this FACT PACK. The data does not contain enough entries to situate it against any particular opponent. The 50th percentile is a broad ranking rather than a head-to-head result.