Performance is also eerily similar on paper despite the half the power limit.
From what I hear NVIDIA are sitting on a lot of Ampere generation workstation desktop cards hence the limited release it will be weird having potentially faster laptops than desktop workstations unless you stump up for the RTX 6000 Ada.
Their pricing for the ampere gen hasn’t seemed to adjust now supply isnt an issue, I guess just little competition from AMD but surely something has to change to force the transition.
The RTX A4000 is also the only ampere or newer single slot card. It hopefully not the last ever, as it can be an issue in workstations though where you might need a pci e slot for extra networking, legacy ports
I do wish nvidia or AMD would just release a proper consumer grade double slot low profile GPU, instead of having to rely on creator/workstation cards. The last one was what... 1050 Ti?
The newest (and most powerful, on paper) low profile consumer cards so far are the GTX 1650 and RX 6400. The former has no RT cores and the latter is pretty much a joke especially on PCIe 3.0, so the market for this kind of cards are definitely stagnating.
Historically Nvidia was using the first letter of the core architecture to distinguish the generation. Like P2000 vs M2000. But I guess they ran into problem with Ampere and Ada share the same first letter. So they went for the USB-IF naming scheme, presenting you RTX A6000 Ada Lovelace.
Nvidia used to use surnames to identify professional cards (exception was Turing architecture),so RTX L6000 (Lovelace) etc would have been absolutely logical avoiding such a mess
Indeed and they are even doing this in the Data Center. So A40 became L40, T4 became L4. Why they insist on making more confusion for workstation users after already scrapping the well known and respected Quadro brand in 2021 I don’t know. Leads to many less educated users buying GeForce by mistake for professional applications. RTX 3000 Ada, RTX 4000 Ada will be even easier to confuse with GeForce RTX 3080. 4080 etc, especially as sometimes the GeForce prefix isn’t prominent on the spec
Does AD102 really have 3 encoders and decoders as listed under RTX 6000? The 4090 has 2 and Nvidia said that Lovelace cards with >12GB get 2 encoders. Just curious if the 4090 has a cut down encoder/decoder system or not.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
15 Comments
Back to Article
meacupla - Tuesday, March 21, 2023 - link
$1250?I am just going to watch reviewers like ETA prime get their hands on this, and never buy it myself.
Ryan Smith - Tuesday, March 21, 2023 - link
The price is comparable to RTX A4000 at launch, so it's not new territory for the ex-Quadro cards.BR1989 - Tuesday, March 21, 2023 - link
Performance is also eerily similar on paper despite the half the power limit.From what I hear NVIDIA are sitting on a lot of Ampere generation workstation desktop cards hence the limited release it will be weird having potentially faster laptops than desktop workstations unless you stump up for the RTX 6000 Ada.
Their pricing for the ampere gen hasn’t seemed to adjust now supply isnt an issue, I guess just little competition from AMD but surely something has to change to force the transition.
The RTX A4000 is also the only ampere or newer single slot card. It hopefully not the last ever, as it can be an issue in workstations though where you might need a pci e slot for extra networking, legacy ports
Oxford Guy - Friday, March 24, 2023 - link
'Their pricing for the ampere gen hasn’t seemed to adjust'Check out what people are trying to sell Volta for. That territory is purely irrational.
meacupla - Tuesday, March 21, 2023 - link
I do wish nvidia or AMD would just release a proper consumer grade double slot low profile GPU, instead of having to rely on creator/workstation cards.The last one was what... 1050 Ti?
ballsystemlord - Tuesday, March 21, 2023 - link
No, an RX550.AMD launched that about 6 months after Nvidia launched the 1050ti.
meacupla - Wednesday, March 22, 2023 - link
Okay, these are both ancient cards by now.logoffon - Wednesday, March 22, 2023 - link
The newest (and most powerful, on paper) low profile consumer cards so far are the GTX 1650 and RX 6400.The former has no RT cores and the latter is pretty much a joke especially on PCIe 3.0, so the market for this kind of cards are definitely stagnating.
ballsystemlord - Thursday, March 23, 2023 - link
The RX6400 is a way to get us to upgrade our MBs. Admit it.TheinsanegamerN - Tuesday, March 21, 2023 - link
I look forward to them being available for $250 in 2 years, so I can finally get a SFF GPU upgrade for my media PCPiotrek54321 - Wednesday, March 22, 2023 - link
That naming scheme is horrible. Why are they using names from what would normally be 5 generations (RTX 2xxx, 3xxx, 4xxx, 5xxx, 6xxx)?erinadreno - Wednesday, March 22, 2023 - link
Historically Nvidia was using the first letter of the core architecture to distinguish the generation. Like P2000 vs M2000. But I guess they ran into problem with Ampere and Ada share the same first letter. So they went for the USB-IF naming scheme, presenting you RTX A6000 Ada Lovelace.nichimen - Wednesday, March 22, 2023 - link
Nvidia used to use surnames to identify professional cards (exception was Turing architecture),so RTX L6000 (Lovelace) etc would have been absolutely logical avoiding such a messAlsw - Wednesday, March 22, 2023 - link
Indeed and they are even doing this in the Data Center. So A40 became L40, T4 became L4. Why they insist on making more confusion for workstation users after already scrapping the well known and respected Quadro brand in 2021 I don’t know. Leads to many less educated users buying GeForce by mistake for professional applications. RTX 3000 Ada, RTX 4000 Ada will be even easier to confuse with GeForce RTX 3080. 4080 etc, especially as sometimes the GeForce prefix isn’t prominent on the specOtritus - Monday, March 27, 2023 - link
Does AD102 really have 3 encoders and decoders as listed under RTX 6000? The 4090 has 2 and Nvidia said that Lovelace cards with >12GB get 2 encoders. Just curious if the 4090 has a cut down encoder/decoder system or not.