NVIDIA RTX 6000 Ada was tested in 3DMark

NVIDIA RTX 6000 Ada was examined in 3DMark
NVIDIA’s new RTX 6000 was examined in 3DMark
Reddit has a thread that includes NVIDIA’s not too long ago launched RTX 6000 ADA graphics card. That is by far probably the most highly effective AD102 GPU variant in the marketplace.
The RTX 6000 ADA is just not a gaming card, nevertheless. NVIDIA has a separate line of workstation fashions designed particularly for this goal. Subsequently, this card truly has a considerably decrease TDP (300W) in comparison with RTX 4090 (450W).
NVIDIA RTX 6000 ADA, Supply: Wholesome-Blood-54
A Redditor claims that the cardboard runs at 85°C with followers round 70%, it is no louder than an RTX A6000 (the predecessor). Though it could appear excessive, it have to be remembered that it is a blower-type cooler and these playing cards are designed to function beneath heavy workloads for a protracted interval.
As a reminder, RTX A6000 ADA has 18176 CUDA cores, so 1792 cores or 10% greater than RTX 4090. Increased core depend and decrease TDP aren’t the one variations. This workstation card is supplied with 48GB of GDDR6 reminiscence with 20 Gbps, so the efficient bandwidth is barely 48 GB/s decrease than RTX 4090.
NVIDIA RTX 6000 ADA, Supply: Wholesome-Blood-54
The one screenshot posted exhibits the cardboard scoring 30518 factors in 3DMark TimeSpy Graphics, which is 15% slower than the RTX 4090. Nonetheless, it might reportedly hit 36K factors earlier than a Home windows replace. That mentioned, 3DMark efficiency is at the moment everywhere, and we do not know if the person was utilizing the most recent drivers.
The 6000 ADA ought to a minimum of match the RTX 4090 benchmarks, however contemplating how the TDP is restricted, it could certainly have bother being quicker in additional demanding workloads. With rumors of a possible RTX 4090 Ti having the identical variety of cores, this will likely clarify why this subsequent flagship gaming GPU can declare near 600W.
Supply: Reddit your @harukaze5719
title_words_as_hashtags]