Desktop graphics card with 2.4 GHz clock is featured in Geekbench


from Valentine Sattler
Intel’s Arc A770 desktop graphics card appeared in the Geekbench database. According to the entry, the model relies on 512 compute units and a 2.4 GHz clock. However, the performance values ​​obtained with it are not so promising.

While Intel’s first Arc laptop GPUs are increasingly appearing in benchmarks, the company is already working on the first corresponding desktop solutions. According to the current state of knowledge, these should hit the market this summer and compete with competition from Nvidia and AMD with up to 512 compute units. The flagship would be called the Intel Arc A780, but full shader extension would be used in the next smaller model, the Arc A770.

Reference for Arc A770

This same second fastest model, the Intel Arc A770, recently appeared in the Geekbench database. So there are no longer just first performance data, but you have details on the technology. However, the information should be taken with caution, after all it could be a pre-production model and not the finished product.

If you believe the entry, the Arc A770 will rely on said 512 compute units and therefore 4,096 shaders. The clock is 2.4 GHz, which would put the model slightly above previous laptop models. On the other hand, the A770M could also voluntarily clock a little lower to make room for the A780M. The specified memory expansion is more difficult to explain: here we are talking about a rather twisted 12.7 GB. However, 16 GB was probably installed on the tested graphics card and the remaining memory was allocated for other purposes.

Also interesting: Intel Arc A350M: According to the test, the smallest laptop GPU is a bit faster than the Geforce MX450

The performance raises other questions. In Geekbench 5’s OpenCL benchmark, the system scores an overall score of 85,585 points, which means that the Arc A770M would quite exactly match the performance level of the Geforce RTX 2070. As for the current generation, the card graphics from Intel would be significantly slower than a Geforce RTX 3060.

However, it should be mentioned once again that these are preliminary benchmarks and should not necessarily reflect final performance. Also, while measured compute performance is important for some tasks, it can deviate significantly from other applications, such as gaming performance. other benchmarks.

Source: Geekbench via Videocardz

Leave a Comment