ATI Radeon 9000 PRO Mac Edition

The ATI Radeon 9000 PRO Mac Edition is an end-of-life AGP 4x GPU from AMD built around the RV250 chip. It belongs to the Radeon R200 (9000) generation and uses the Rage 7 architecture, with TSMC manufacturing the die on a 150 nm process. The transistor count is 36 million on a 97 mm² die, giving a density of 371.1K / mm². Its release date is 2004-03-13, and the product lineage places it after the Radeon R100 and before the Radeon R300. The database currently records a 50th percentile across all GPUs, but the benchmark array is empty and the average benchmark score is 0.

How It Compares

The nearestRivals array for this product is empty. There are, therefore, no rival names, no comparison scores, and no deltaPct values to interpret in this section. The only positional data is the 50th percentile, which places the card at the midpoint of the database's GPU ranking. Being at the median suggests a middle-of-pack product rather than a high-end or low-end outlier, though this interpretation is limited because no per-workload scores are present to confirm it.

The predecessor and successor fields are the other relational anchors. The Radeon R100 is listed as predecessor and the Radeon R300 as successor; these are chronological lineage relationships, not measured performance rivals. Without nearestRivals entries, a rival-by-rival comparison is not supported by the fact pack.

Ray Tracing and Feature Set

The RT core and tensor core fields are both null in the fact pack, so no hardware ray tracing or tensor acceleration capacity is recorded for this GPU. The API surface is likewise tied to an earlier feature era: DirectX 8.1 and OpenGL 1.4 are listed, while Vulkan is not present. This combination places the card's feature set before modern ray-traced and tensor-heavy workloads.

The shading unit count is not provided. What is provided are 4 TMUs and 4 ROPs. Those units align with a conventional rasterization pipeline. The pixel rate is 1.100 GPixel/s and the texture rate is 1.100 GTexel/s, both of which are modest numeric ceilings. The absence of Vulkan, RT cores, and tensor cores means there is no data path in this record for ray tracing or AI-accelerated features.

Memory Subsystem

The memory configuration consists of 64 MB of DDR RAM. The bus is 128 bits wide, and the memory clock is listed at 275 MHz, with an effective data rate of 550 Mbps. Multiplying those factors gives a bandwidth of 8.800 GB/s. For high resolutions, the 64 MB capacity is the most immediate limit: every framebuffer, texture set, and depth buffer must fit into that space, and demanding scenes can exceed it quickly.

A 128-bit bus for DDR at 275 MHz produces 8.800 GB/s; that is the hard bandwidth ceiling for texture streaming and framebuffer writes. With 4 ROPs and a 1.100 GPixel/s pixel rate, fill rate can also become the binding constraint at higher resolutions. The memory subsystem is coherent for its time, but it is not built for large, high-resolution working sets.

FAQ

Q: What chip and architecture are used?

A: The card is based on the RV250 chip from AMD, uses the Rage 7 architecture, and belongs to the Radeon R200 (9000) generation. It is fabricated by TSMC on a 150 nm process, with 36 million transistors and a 97 mm² die.

Q: Which APIs are supported?

A: The API list includes DirectX 8.1 and OpenGL 1.4. Vulkan is not listed in the fact pack.

Q: How much memory does it have, and what is the bandwidth?

A: It has 64 MB of DDR memory on a 128-bit bus. The memory clock is 275 MHz, with 550 Mbps effective data rate, and the resulting bandwidth is 8.800 GB/s.

Q: What power supply and connectors are required?

A: The TDP is 28 W, the suggested PSU is 200 W, and there are no power connectors. The card is single-slot.

Q: What display outputs are present?

A: Two DVI outputs are listed.

Q: What benchmark data is available for this card?

A: The benchmarks array is empty, the average benchmark score is 0, and the nearestRivals array is empty. The only ranking field is a 50th percentile across all GPUs.

Benchmark Performance

Because nearestRivals is empty, there are no exact percentage deltas to report versus other GPUs. The fact pack does not contain rival names, scores, or deltaPct values. The average benchmark score is 0, which represents the absence of populated benchmark results, not a measured zero-performance result. The 50th percentile across all GPUs is the only aggregate ranking available.

The card's quantitative rendering limits come from the fill-rate fields: 1.100 GPixel/s and 1.100 GTexel/s. These are derived from the 4 ROPs and 4 TMUs, and they act as ceilings for pixel and texture work. Without actual benchmark workloads, the data cannot support statements such as "ahead by a certain percentage" or "beats a named rival." The record simply does not contain the inputs for that analysis.

Power and Cooling

The GPU is rated at 28 W TDP. The recommended PSU for the system is 200 W. No auxiliary power connectors are listed, so the card does not need external power cabling. Its physical layout is single-slot, which makes it suitable for chassis with limited expansion width.

Manufactured on a 150 nm process with 36 million transistors, the card's thermal budget is low enough that the listed power requirements are modest. The database lists no dimensions, so the single-slot field is the only physical geometry available. The low TDP and 200 W PSU figure mean system planning can assume a low-power AGP 4x board.

Who Should Consider It

This card is suited to older AGP 4x systems needing dual DVI outputs. The absence of power connectors and the 28 W TDP make it a low-power candidate for chassis where additional cables are undesirable. The 200 W suggested PSU is a modest system-level requirement.

Software support matters: DirectX 8.1 and OpenGL 1.4 mean the card is best matched to applications from that API generation. Vulkan is absent, and RT/tensor core fields are null, so modern, API-heavy workloads are not indicated. With 64 MB of DDR and 8.800 GB/s bandwidth, lower resolutions and reduced texture settings are the appropriate envelope. At higher resolutions, the 4 ROPs and 1.100 GPixel/s pixel rate will constrain fill, and 64 MB will constrain capacity.

The 50th percentile position suggests a median standing in the database, but because no benchmark scores are attached, this should be read cautiously. Users with legacy AGP 4x motherboards, a 200 W class PSU, and a need for 2x DVI are the audience the specification sheet points to.