ATI Radeon SDR PCI

The ATI Radeon SDR PCI is a Rage 6 architecture graphics card from AMD, built on TSMC's 180 nm process with 30 million transistors on a 115 mm² die, giving a transistor density of 260.9K per mm². Released on 2000-05-31, the card occupies the 50th percentile in the benchmark database, with an average benchmark score of 0. It uses the PCI bus interface, comes in a single-slot form factor, and its production status is end-of-life.

Power and Cooling

The Radeon SDR PCI draws a TDP of 23 W. That figure places it well within the reach of standard desktop power supplies of its generation. The data shows no power connectors on the card — power is drawn entirely from the PCI slot. AMD's suggested PSU rating is 200 W, which is a conservative recommendation for a system with this card installed. The single-slot cooler is adequate for the 23 W thermal envelope; no auxiliary cooling is required. The card's physical length is 165 mm, or 6.5 inches, which allows it to fit in most compact chassis of the era. The 180 nm process node and the 30 million transistor count are the key contributors to the low power draw. The absence of a supplementary power connector simplifies installation and reduces cable clutter in legacy systems. The 200 W suggestion accounts for the rest of the system's components, not just the graphics card, and the data indicates that even modest power supplies can handle this configuration. For a card of this class, the power story is straightforward: low draw, no extra cables, and a wide margin of compatibility with period-appropriate power supplies.

Memory Subsystem

The memory subsystem is built around 32 MB of SDR memory, clocked at 166 MHz, and connected via a 128-bit bus. The resulting bandwidth is 2.656 GB/s. The 128-bit bus is a significant asset; it allows the SDR memory to achieve bandwidth that would otherwise require a faster memory type. However, SDR memory transfers data once per clock cycle, unlike DDR which transfers twice per cycle. At 166 MHz, the 128-bit bus yields 2.656 GB/s, which is the figure that all rendering operations must share. For high-resolution scenarios, this bandwidth is the primary constraint. The 32 MB capacity further limits the working set for textures and framebuffer data. When resolution and detail settings increase, the demand for texture reads and framebuffer writes grows, and the 2.656 GB/s pipe becomes the bottleneck. The pixel rate of 332.0 MPixel/s and the texture rate of 996.0 MTexel/s are aligned with what the memory bandwidth can feed. In practical terms, the card is best suited to lower-detail settings at higher resolutions, or higher-detail settings at more modest resolutions — though the data does not specify exact resolution figures. The 128-bit bus width is a point in its favor, as it provides a balanced path for both read and write operations. The memory type being SDR rather than DDR means that the effective transfer rate is capped at one transfer per clock, which is the fundamental limit behind the 2.656 GB/s figure.

Ray Tracing and Feature Set

The Rage 6 architecture does not include ray tracing cores or tensor cores; those entries are absent from the specification. This is a fixed-function design from the DirectX 7.0 era. The card supports DirectX 7.0 and OpenGL 1.3, which defines the software interface available to applications. With 6 texture mapping units and 2 render output units, the card delivers a texture rate of 996.0 MTexel/s and a pixel rate of 332.0 MPixel/s. The 2 ROPs limit fill-rate-heavy workloads, particularly at higher resolutions where more pixels must be written per frame. The absence of dedicated ray tracing hardware means that any such effects would have to be computed on the CPU, which is impractical for real-time use. The feature set is entirely rasterization-based. The display outputs are 1x VGA and 1x S-Video, which cover analog monitors and television output of the period. There is no support for newer API features such as Vulkan, which is not listed in the specification. The card's capabilities are defined by the DirectX 7.0 and OpenGL 1.3 feature levels, and the data shows no extension beyond those. The 6 TMUs and 2 ROPs are the only fixed-function units available for geometry and pixel processing, and their rates — 996.0 MTexel/s and 332.0 MPixel/s respectively — represent the ceiling for texture and pixel throughput.

How It Compares

The nearestRivals field in the data is empty, meaning no direct rival benchmark scores are available for this card. The only comparative metric is the 50th percentile ranking across all GPUs in the database. This places the Radeon SDR PCI exactly at the median of the performance distribution — half of the recorded GPUs are faster, and half are slower. The average benchmark score of 0 indicates that the card has not been exercised by the current benchmark suite, so the percentile is a historical placement rather than a live measurement. The predecessor is the Rage 4 and the successor is the Radeon R100, but no scores are recorded for either in this data, so a direct generational comparison cannot be quantified. The empty rival list means that the card's position must be interpreted through the percentile alone. At the 50th percentile, it is neither a high-end nor a low-end part; it sits in the middle of the database's distribution. Without rival data, further ranking is not possible from the available facts. The percentile is the single reference point, and it suggests a balanced, mid-pack positioning that aligns with the card's modest but not trivial specifications.

Who Should Consider It

The 50th percentile placement indicates a mid-range position in the overall GPU landscape. However, the specifications tell a more specific story. The 32 MB memory capacity and 2.656 GB/s bandwidth are modest by any standard, and they constrain the card to lower-detail settings. The 23 W TDP and PCI bus interface make the card an easy fit for legacy systems where power delivery and slot compatibility are the primary concerns. The DirectX 7.0 API support limits the software library to titles from that generation; modern games that require newer API versions will not run. The end-of-life production status means the card is no longer manufactured. For users building a retro system or repairing a vintage machine, the Radeon SDR PCI offers a functional option with a known performance envelope. The 332.0 MPixel/s pixel rate and 996.0 MTexel/s texture rate provide a baseline for playable settings in era-appropriate titles. Users who need high-resolution, high-detail rendering would find the memory bandwidth and capacity insufficient. The card is not suited to ray-traced workloads, as it lacks the dedicated hardware. The data supports a clear recommendation: this card belongs in a period-correct build, not a modern workstation.

FAQ

Q: How much memory does the ATI Radeon SDR PCI have?

A: It has 32 MB of SDR memory on a 128-bit bus.

Q: Does this card require a power connector?

A: No. The specification lists no power connectors, and the suggested PSU is 200 W.

Q: What is the TDP of this card?

A: The TDP is 23 W, with a single-slot cooler.

Q: What API versions does it support?

A: It supports DirectX 7.0 and OpenGL 1.3.

Q: What is the memory bandwidth?

A: The memory bandwidth is 2.656 GB/s, derived from a 166 MHz memory clock on a 128-bit bus.

Q: Is this card still in production?

A: No. The production status is end-of-life, and it was released on 2000-05-31.