ATI Mobility FireGL T2e

The ATI Mobility FireGL T2e is a mobile workstation GPU built on the R300 architecture. AMD fabricated the M11 chip at TSMC on a 130 nm process, packing 76 million transistors into a 92 mm² die with a transistor density of 826.1K per mm². Released on 2004-07-31 and now end-of-life, the card carries 128 MB of DDR memory on a 128-bit bus with 7.200 GB/s of bandwidth. The database records a 50th-percentile standing among all GPUs, placing it exactly at the median of tracked parts. With 4 TMUs and 4 ROPs, it delivers a pixel rate of 1.800 GPixel/s and a texture rate of 1.800 GTexel/s. The API support covers DirectX 9.0 and OpenGL 2.0, with no Vulkan, no RT cores, and no tensor cores.

Who Should Consider It

Benchmark results place this card at the 50th percentile of all GPUs in the database, a median position that tells a clear story: half of tracked parts outperform it and half fall behind. For a mobile workstation part from the R300 generation, that standing translates to entry-level 3D workloads rather than high-end rendering. The 128 MB frame buffer and 7.200 GB/s bandwidth are the binding constraints. At lower resolutions and with reduced texture detail, the card can keep pace with the modest fill rates of 1.800 GPixel/s and 1.800 GTexel/s; at higher resolutions, the memory capacity becomes the limiting factor. Users running legacy DirectX 9.0 applications or fixed-function OpenGL 2.0 workloads that fit within 128 MB will find a usable, if dated, tool. Anyone expecting modern high-resolution or shader-heavy performance should look at more recent parts. The card is end-of-life, so its practical audience is limited to retro environments, legacy industrial software, or collectors of early 2000s mobile workstations. The 50th-percentile rank suggests it is neither a standout nor a straggler — it is the definition of a mid-pack mobile GPU.

Ray Tracing and Feature Set

The R300 architecture predates dedicated ray tracing hardware, and the FACT PACK confirms this: no RT cores and no tensor cores are listed for the Mobility FireGL T2e. The API surface is DirectX 9.0 and OpenGL 2.0, with no Vulkan support. This means the card cannot accelerate any ray-traced effects, denoising, or AI-based upscaling — features that rely on tensor cores in modern parts. The feature set is purely rasterization-oriented: 4 TMUs and 4 ROPs drive the fixed-function pipeline that defined early 2000s graphics. OpenGL 2.0 does bring programmable shaders of that era, so developers could write vertex and fragment shaders within the limits of DirectX 9.0. But there is no geometry shader stage, no tessellation, and no compute shader support — those arrived in later API revisions. In practical terms, the Mobility FireGL T2e is a raster-only device. Any workload that depends on modern API features — ray tracing, mesh shaders, or tensor-core acceleration — will not run on this hardware. Its feature set is a snapshot of 2004-era mobile workstation capability, and it should be judged as such.

How It Compares

The FACT PACK's nearestRivals field is empty, so there are no direct rival scores or deltaPct values to cite. The only comparative anchor is the 50th-percentile ranking against all GPUs in the database. That median position means the card outperforms roughly half of all tracked GPUs and trails the other half. Without rival data, the comparison must rest on the card's own throughput metrics: a pixel rate of 1.800 GPixel/s and a texture rate of 1.800 GTexel/s, fed by 7.200 GB/s of memory bandwidth. These figures are internally consistent — the memory bandwidth is sufficient to sustain the fill rates, so the card is not obviously bottlenecked at its own design point. The absence of recorded rivals is itself informative: it suggests the database does not track direct contemporaries for this mobile workstation part, or that its performance class was considered too niche for head-to-head comparison. In the broader database, a 50th-percentile GPU is the baseline against which faster and slower parts are measured. The Mobility FireGL T2e is that baseline. No percentage deltas can be computed because no rival scores exist in the record.

FAQ

Q: How much memory does the ATI Mobility FireGL T2e have?

A: It has 128 MB of DDR memory on a 128-bit bus, providing 7.200 GB/s of bandwidth.

Q: What API versions does it support?

A: It supports DirectX 9.0 and OpenGL 2.0. Vulkan is not supported.

Q: What is the memory clock speed?

A: The memory runs at 225 MHz, with 450 Mbps effective data rate.

Q: What process node and foundry are used?

A: TSMC fabricates the chip on a 130 nm process, with 76 million transistors on a 92 mm² die.

Q: What is the successor to this GPU?

A: The successor is the FirePro Mobility line.

Q: When was it released and what is its production status?

A: It was released on 2004-07-31 and is now end-of-life.

Q: Does it have ray tracing or tensor cores?

A: No — the FACT PACK lists no RT cores and no tensor cores.

Memory Subsystem

The memory subsystem consists of 128 MB of DDR memory on a 128-bit bus. The memory clock is 225 MHz with a 450 Mbps effective data rate, which yields a bandwidth of 7.200 GB/s. For high resolutions, the 128 MB capacity is the primary constraint. Texture-heavy scenes will exceed the frame buffer quickly, forcing the application to reduce detail or resolution. The 128-bit bus width is moderate for the era, and the resulting 7.200 GB/s bandwidth aligns closely with the card's pixel and texture rates of 1.800 GPixel/s and 1.800 GTexel/s. This balance means the memory subsystem is not the bottleneck at the card's native fill rates — the raster pipes and the memory feed each other adequately. However, the small capacity means that any workload exceeding 128 MB of working set will see severe performance degradation as data spills to system memory over the AGP 4x interface. The AGP 4x bus itself is another constraint; it caps the transfer rate between the card and the host, which matters when the local frame buffer is exhausted. For a 2004-era mobile workstation, this configuration was reasonable, but it is not suited to modern high-resolution textures or large datasets.

Power and Cooling

The FACT PACK records no TDP, no suggested PSU, no power connector requirements, and no slot width for the Mobility FireGL T2e. As a result, any specific wattage or cooler recommendation would be unsupported speculation. The absence of documented power connectors suggests that the card may draw its power through the AGP 4x slot alone, but this is an inference from missing fields rather than a recorded fact. Similarly, the thermal design is unquantified; the card's 130 nm process and 76 million transistors imply a certain power envelope, but no number is given. The bus interface is AGP 4x, which is the only connectivity fact available. For cooling, a capable air cooler would be the qualitative expectation for a mobile workstation GPU of this vintage, but no cooler size, fan count, or thermal rating is listed. System integrators should treat the power and cooling requirements as undocumented and verify them from the original vendor documentation. The end-of-life status further complicates modern deployment, as replacement coolers and power supplies are not specified in the database.

Benchmark Performance

The benchmark array for the Mobility FireGL T2e is empty, and the average benchmark score is recorded as zero. This means the database holds no direct performance measurements for this part. The only quantitative performance reference is the 50th-percentile ranking, which places the card at the median of all GPUs tracked. The raw throughput figures define the rasterization ceiling: a pixel rate of 1.800 GPixel/s and a texture rate of 1.800 GTexel/s, produced by 4 TMUs and 4 ROPs. The memory bandwidth of 7.200 GB/s is exactly enough to feed those rates — at 128-bit bus width and 450 Mbps effective, the math is consistent. With no nearest rivals and no recorded scores, percentage deltas cannot be computed. The percentile is the sole comparative metric, and it indicates a median performer. In practical terms, the card will handle early DirectX 9.0 workloads at its native fill rates, but any scene that stresses the pixel or texture pipes beyond 1.800 GPixel/s or 1.800 GTexel/s will cause frame rate drops. The zero benchmark score should not be read as zero performance; it reflects the absence of recorded data, not the absence of capability. The 50th percentile remains the best available signal: a mid-pack mobile workstation GPU that is neither exceptional nor deficient.