As an aside, I am sure most people do not know this as well. The reason it shows at 44GB of RAM is because of ECC. ECC is on by default, it uses one chip for parity data, so each chip on that board must be 4GB. My K80s in ECC mode go from 12GB to 11GB. In the Nvidia control panel, you can disable ECC(requires a reboot), if you need the extra 4GB.
Imagine if on the servers when you put 32GB DDR-4 you get only 30, because of the ECC 🙂. So NVidia is *lying* that there are 48GB ECC - there are 44GB ECC.
That would be an interesting comparison, and probably something outside of the range of what I could setup right now. My guess is that it would be similar if you do not need >24GB RAM on the GPU. For something that parallel-izes really well (GAN training comes to mind), two 3090s would likely outperform a single A6000. Of course, realize that 24GB on two 3090s can't do all of the same things as 48GB on a single A6000. Generally, you are mirroring training data and parameters. Cooling two 3090s is going to be more challenging than cooling a single A6000, or even two A6000s.
I really appreciate NVidia sending all these GPUs to Data Science and deep learning TH-camrs. Makes me really happy. But will it also do the same for those who write deep learning tutorials? Like I have my own deep learning/machine learning website where I write articles and tutorials. I require GPU on a regular basis, but there is only so much I can do with my GTX 1060 laptop. I really want to amp up my tutorial quality. Hope NVidia does the same for those who write tutorials as well. Great video as always.
So, how you attract a vendor's attention for this sort of thing, is to publish your material to social media. Tag the vendor when you post it, this is how I attracted some attention. I am far from a mega-channel.
@@HeatonResearch Wow! really thank you for the advice. Will surely try to implement it. And regarding your channel, it is one of the best blends of knowledge about deep learning hardware and software that I have ever seen. You show us the hardware, you go into detail about its specification, you even show how to install it in a machine. And your code explanations are just as good. I hope that your channel really takes off to millions of subscribers.
At the time that was infeasible, with a max of 2GB per GDDR6 chip and maximum of 24 memory chips (12 on front, 12 on back) on the PCB. But with stacked GDDR6W (4GB per chip), we might finally see 96GB cards soon!
HI Jeff, and thank you for your all quality videos. You spoke about comparaison between RTX 3090 and A6000, it is possible to have you opinion about the Nvidia A30 option ? The price is really close to the A6000. Thanks for all from France ;)
The A30 is a cut-down A100 with only 24GB, with a GA100 die inside. It can do FP64 at 1:2 rate to FP32. The A6000 48GB and 3090 24GB are both GA102 and mostly incapable of FP64 (1:64 rate to FP32).
Thanks so much for such awesome content. A humble request, would you please highlight whether the performance is worth the expense. Would it be better to buy 2 x 3090s? Perhaps indicate from your experience whether the unlocked features as well as the ECC ram on the card affect inference to the extent that 3090s aren't appropriate for DataScience.
That would be an interesting comparison, and probably something outside of the range of what I could setup right now. My guess is that it would be similar if you do not need >24GB RAM on the GPU. For something that parallel-izes really well (GAN training comes to mind), two 3090s would likely outperform a single A6000. Of course, realize that 24GB on two 3090s can't do all of the same things as 48GB on a single A6000. Generally, you are mirroring training data and parameters. Cooling two 3090s is going to be more challenging than cooling a single A6000, or even two A6000s.
@@HeatonResearch I can add a bit about that. Using two GPUs, the easiest way to achieve parallelism is indeed data parallelism (i.e. Pytorch's DataParallel). In this way, data batches are not mirrored, and each gpu gets its own half of the current minibatch. Other than that, bear in mind that there are 3090s specifically made for multi-gpu configurations. Actually you can tightly stack four of them in a desktop, and still get zero cooling issues since they blow all the hot air out of the case: www.pugetsystems.com/labs/articles/Quad-GeForce-RTX-3090-in-a-desktop---Does-it-work-1935/ I got two of them, and the maximum temp I get is barely 71-72C for the upper card.
@@weylandsmith5924 Hey Weyland, would you recommend a 2x3090 over 1xA6000? I am looking at several builds, but getting some mixed opinion esp. that the performance increase giong 2x3090 may not be that huge (although the CUDA cores are so many more)
You should have left the Titan in the next slot. I'm fairly sure the Studio drivers are the same for both cards. For example I use a RTX 3060 and a RTX 2070 Super. They work perfectly together for video editing
A few more years and I'll be able to afford this card for my ESXi home lab server since it supports vGPU profiles and has a ton of VRAM. Might get the A5000 for now upgrade from the Tesla P4.
Fantastic video. Thank you very much for sharing your experience. I am getting a new workstation with the main purpose of 8K video editing. Did you see faster rendering of your TH-cam video with this card?
I do not do 8K at this point but for HD and 4K, premiere pro loves the new card. Have not formally benchmarked it though, does feel faster. This will only improve as software better embraces Ampere.
@@HeatonResearch could you recommend a retailer and/or affiliate code for a retailers for this card? Can you also suggest any that do educational discounts with Nvidia?
@@MatthewRalston89 There are indeed EDU discounts, but you have to be a teacher, not a student. For example, delta computers offers the A6000 for 3700eur with the EDU discount. But then you can buy two 3090s with such money, even without any discount.
Hey! thanks for an overview video! May I ask, what are the noise levels in full load? I would Assume very loud since it's like 3090 with blower style and 300W, but I see you have it right next to the monitor, so I'm curious how much dB does it produce? Thanks!
I have my recording setup/mics specifically tuned to eliminate noise, such as the fan or my dog snoring under the desk. Measured with a highly scientific iPhone 😊, the room, with nothing running is 45db and 60db with the A6000 running full fan.
Nice anti-static handling technique. I highly recommend it for all nvidia cards. Also nvidia cards may have a massive mis-match between compute and memory bandwidth for the newer and faster NN algorithms. For the NN segment of the market they face huge re-engineering costs. However I'm sure they have the money. Anyway it's not their fault. It's down to the broken foundations of NN research. The failure even to apply the scientific method to the dot product and other simple aspects of NNs.
Yeh, they don't even know how to use the dot product as a general associative memory by meeting the demands of the variance equation. Ie. Random projection, non-linear function, dot product. Terrible basic research. The NN books start in the middle and work forward. No one looked back to check the basics and challange assumptions.⚠⚠⚠
If you are wondering what the actual box looks like, if you get this card through more traditional retail, www.leadtek.com/eng/products/workstation_graphics(2)/NVIDIA_Quadro_RTX_A6000(30893)/detail
@Jeff Heaton, Hey man just a word of advice. Never daisy chain power connectors to your rtx cards. Yeah I know psu supply them but it's their way of cutting costs. Always use separate 8 pin connects for each plug in, if that makes sense. Yeah most won't care and do it daisy chaining, but if you're using it for high end rendering, ai simulations, neural networks, or in my case 4x 3090 FE's for mining, separate connections EACH to pardon my French, 'Cover your @$$'... See ya im outta here~
look world gloves!! yup! 4 real !! nice presentation Jeff! popcorn engaged! the Angel on his shoulder says: hand with care!! the devil on the other shoulder, yells out. nail this card 2 the floor Jeff!! woohoo!! I'm like thinking, mastercard ''will'' cover & warranty this product!! though I ''think'' nVidia has a lifetime guarantee. but I'm not 4 sure!! good luck!!
@@HeatonResearch nVidia has just released the A4000 & A5000 😃 The A4000 looks interesting with 16GB ram at only USD999, but has no nvLink 😥 btw how loud is the blower style cooler when running 100% for several minutes? I have heard they are much louder than regular cards?
Running at 80% it is about 60 decibels, as measured by my iPhone. By contrast, a silent room in my house is 40. In an office setting, you would hear it for sure, but not overly obnoxious (at least to me).
We're here for the seeing what the card can do. Why are you wasting 5 minutes just building the computer? There are other channels dedicated to pc builds.
I've used both, and yes the A6000 is faster. Same memory. Right now the A6000 fills the same top spot in the professional workstation GPU line that the Quadro RTX 8000 filled. I am not sure if NVIDIA is holding the "A8000" name for an even more powerful card in this line.
I think that is a real question if you do not need the memory. I don't have the hardware to test that. I am looking forward to some 3090 benchmarks myself.
They did not exist at the time I made this video. I will probably cover them in a future video. I do have videos on cheaper cards, such as: th-cam.com/video/CPAmiRkP3U0/w-d-xo.html
@@HeatonResearch Thank you, would be good if you can do a video to supplement the latest RTX A0000 series, on the RTX A4000 and the RTX5000 seen as you've done one on the RTX6000. These are obviously the more affordable and latest Nvidia GPU's that most of us would be interested learning more about specifically and perhaps some demos on their performance...As the RTX A6000 is out of our budget....and most of us will never see the light of day actually using..... probably a good idea on the more affordable versions of its line. Thank you. Great video Jeff, keep up the great work
i dont care what anyone says, blowers have always been the best for my system... such a pity this design has been pushed out. by far the best looking gpu... every other design has been hot garbage... too gamerish, screams im only 4 foot 5 but i have 22 inch tires with 8 wheel drive that can rock climb with tower like suspension cause it looks "kewl" with tons of RGB
Oh and by the way... Yes, it will run Cyberpunk 2077 :-) instagram.com/p/CLAwFnyhN4p/
Jhonny will be happy
But will it run cyberpunk 2077 at ultra 4K with anti-aliasing? Mwahahahaha!
How many fps?
As an aside, I am sure most people do not know this as well. The reason it shows at 44GB of RAM is because of ECC. ECC is on by default, it uses one chip for parity data, so each chip on that board must be 4GB. My K80s in ECC mode go from 12GB to 11GB. In the Nvidia control panel, you can disable ECC(requires a reboot), if you need the extra 4GB.
Imagine if on the servers when you put 32GB DDR-4 you get only 30, because of the ECC 🙂. So NVidia is *lying* that there are 48GB ECC - there are 44GB ECC.
I would like benchmarks for the A6000 vs a 3090 on the same training set.
Same
That would be an interesting comparison, and probably something outside of the range of what I could setup right now. My guess is that it would be similar if you do not need >24GB RAM on the GPU. For something that parallel-izes really well (GAN training comes to mind), two 3090s would likely outperform a single A6000. Of course, realize that 24GB on two 3090s can't do all of the same things as 48GB on a single A6000. Generally, you are mirroring training data and parameters. Cooling two 3090s is going to be more challenging than cooling a single A6000, or even two A6000s.
@@HeatonResearch yep, 3090 is so big and costing 350w power
Jeff, I got a workstation with two 3090s. If you want, we can run benchmarks and compare them. Just give me the benchmark notebooks and I'll run them.
@@weylandsmith5924 u lucky guy
What a beauty the "professor's" RTX A6000 is. The new bling!
Indeed! Thanks!
I really appreciate NVidia sending all these GPUs to Data Science and deep learning TH-camrs. Makes me really happy. But will it also do the same for those who write deep learning tutorials? Like I have my own deep learning/machine learning website where I write articles and tutorials. I require GPU on a regular basis, but there is only so much I can do with my GTX 1060 laptop. I really want to amp up my tutorial quality. Hope NVidia does the same for those who write tutorials as well.
Great video as always.
So, how you attract a vendor's attention for this sort of thing, is to publish your material to social media. Tag the vendor when you post it, this is how I attracted some attention. I am far from a mega-channel.
@@HeatonResearch Wow! really thank you for the advice. Will surely try to implement it. And regarding your channel, it is one of the best blends of knowledge about deep learning hardware and software that I have ever seen. You show us the hardware, you go into detail about its specification, you even show how to install it in a machine. And your code explanations are just as good. I hope that your channel really takes off to millions of subscribers.
@@sovitrath4735 🤦
waiting for nvidia to release an RTX A8000. I bet that would either be a 64GB or 96GB vram graphics card.
At the time that was infeasible, with a max of 2GB per GDDR6 chip and maximum of 24 memory chips (12 on front, 12 on back) on the PCB. But with stacked GDDR6W (4GB per chip), we might finally see 96GB cards soon!
wow you are so lucky my guy, hope you enjoy that one day ill get one. ONE DAY lol
👆👆 I GOT 3080 WITH A FAIR PRICE FROM THEM
And I was so happy with my RTX 3070 :-)
Cries in integra
@@summiTHVH Drunkenly sobs in Windows Soviet.
HI Jeff, and thank you for your all quality videos. You spoke about comparaison between RTX 3090 and A6000, it is possible to have you opinion about the Nvidia A30 option ? The price is really close to the A6000. Thanks for all from France ;)
The A30 is a cut-down A100 with only 24GB, with a GA100 die inside. It can do FP64 at 1:2 rate to FP32.
The A6000 48GB and 3090 24GB are both GA102 and mostly incapable of FP64 (1:64 rate to FP32).
4:44 - They're DisplayPort 1.4 ports, not DVI. DVI is pretty much dead/legacy, and is a larger white connector.
I HAVE BEEN WAITING FOR THIS!!!!!!!
Thanks so much for such awesome content. A humble request, would you please highlight whether the performance is worth the expense. Would it be better to buy 2 x 3090s? Perhaps indicate from your experience whether the unlocked features as well as the ECC ram on the card affect inference to the extent that 3090s aren't appropriate for DataScience.
That would be an interesting comparison, and probably something outside of the range of what I could setup right now. My guess is that it would be similar if you do not need >24GB RAM on the GPU. For something that parallel-izes really well (GAN training comes to mind), two 3090s would likely outperform a single A6000. Of course, realize that 24GB on two 3090s can't do all of the same things as 48GB on a single A6000. Generally, you are mirroring training data and parameters. Cooling two 3090s is going to be more challenging than cooling a single A6000, or even two A6000s.
@@HeatonResearch I can add a bit about that. Using two GPUs, the easiest way to achieve parallelism is indeed data parallelism (i.e. Pytorch's DataParallel). In this way, data batches are not mirrored, and each gpu gets its own half of the current minibatch.
Other than that, bear in mind that there are 3090s specifically made for multi-gpu configurations. Actually you can tightly stack four of them in a desktop, and still get zero cooling issues since they blow all the hot air out of the case: www.pugetsystems.com/labs/articles/Quad-GeForce-RTX-3090-in-a-desktop---Does-it-work-1935/
I got two of them, and the maximum temp I get is barely 71-72C for the upper card.
@@weylandsmith5924 Hey Weyland, would you recommend a 2x3090 over 1xA6000? I am looking at several builds, but getting some mixed opinion esp. that the performance increase giong 2x3090 may not be that huge (although the CUDA cores are so many more)
You should have left the Titan in the next slot. I'm fairly sure the Studio drivers are the same for both cards. For example I use a RTX 3060 and a RTX 2070 Super. They work perfectly together for video editing
👆 THEY ARE BEST HIT UP FOR A GOOD PRICE
Aah I just ordered a GPU server for my team with dual A100s. Was looking for A40 and A6000 but Dell couldn't validate them in time :(
I will be comparing A100 to A6000 soon, that is on the list. Sounds like I nice system, though.
@@HeatonResearch Interesting.. Will they provide a PCI-express A100?
Why do I need 10,000 cores?
Train a Machine Learning model in hours vs yrs
Train an Ai in years vs decades.
Render a Blender frame in under 10 seconds! Now try making a 5 minute movie (30x60x5 frames) ...
Bruh..... give varying depth Random ReLU feature space expansion a chance........
Have you done any speed tests comparing the A6000 to the 8000?
👆GET FROM THEM FOR A FAIR PRICE
Amazing, this gpu is what dreams are made off xD
I have been dreaming about it lately. 😲😴
Really love your videos! It feels like your uncle teaching you all the cool tech stuff. Cheers and Happy Holidays!
Thanks for sharing Jeff!
Hi Jeff, can you please do a model comparison between a 3090 and the A6000 please in terms of speed?
Woahhh !!!!!! The great Jeff Heaton 🤚
Thanks! And thank you NVIDIA!
@@HeatonResearch sir can you do one for AMD ... For ROCm library .. there is really minimal information about it
You never clarified if the A6000 is a turbo/blower GPU, or not? I would be really interested to know.
You’re the best Jeff!
Seems like jeff's low-key the relaxing end
Damn. Just, damn. I just found a new shade of green for my jealousy, Jeff. :-)
A few more years and I'll be able to afford this card for my ESXi home lab server since it supports vGPU profiles and has a ton of VRAM. Might get the A5000 for now upgrade from the Tesla P4.
still relevant^.^ thank you!
I use two of these as gpu accelerators in CFD simulations in ANSYS FLUENT.
Fantastic video. Thank you very much for sharing your experience. I am getting a new workstation with the main purpose of 8K video editing. Did you see faster rendering of your TH-cam video with this card?
I do not do 8K at this point but for HD and 4K, premiere pro loves the new card. Have not formally benchmarked it though, does feel faster. This will only improve as software better embraces Ampere.
@@HeatonResearch Thank you very much.
Great overview!! Just bought four myself :) Did you manage to find a PSU cable with two PCIe going to one CPU 8-pin connector?
Ah, very cool! I ordered one and am waiting on that currently. Once I am sure I won't be occasionally swapping the TITAN RTX back in, I will switch.
All ways best as your videos series.
Thanks!
I still remember my first video card purchase that was ATi Radeon 512MB at that time I thought was a lot look now wow!
Please run some scientific benchmarks on it! A few batch tests and use cases with different resolutions against other gpus.
ı want to this D: . Hey NVIDIA or Investors I am creating a start up. We need that ( 4 a6000 :( )
Cloud use of A100 might be a good option.
@@HeatonResearch could you recommend a retailer and/or affiliate code for a retailers for this card? Can you also suggest any that do educational discounts with Nvidia?
@@MatthewRalston89 There are indeed EDU discounts, but you have to be a teacher, not a student. For example, delta computers offers the A6000 for 3700eur with the EDU discount. But then you can buy two 3090s with such money, even without any discount.
the GPU is not for gaming, but making the game :D
Mom, i need this for my online schooll... 😗
Need to just get it on the school supply lists.
The best explication 😀😀😀
Thank you so much 😀
Can you add more A6000s to it? does the motherboard support it?
Put the Titan on the second slot and utilize both 😉
Can you please review the A4000?
👆GET FROM THEM FOR A FAIR PRICE
Im gonna buy, This, Since Prices, are so out of hand, anyways.
👆THEY GOT DOPE PC HIT UP
Would you please compare it with Intel Compute Stick 2?
"This is an expensive car ... I mean card" Titan RTX looks like a work of art compared to that black rectangular box.
You sir might be the luckiest guy in the world if not one of the luckiest.
Thanks! Yes I am quite thankful to NVIDIA for this.
Hey! thanks for an overview video! May I ask, what are the noise levels in full load? I would Assume very loud since it's like 3090 with blower style and 300W, but I see you have it right next to the monitor, so I'm curious how much dB does it produce?
Thanks!
I have my recording setup/mics specifically tuned to eliminate noise, such as the fan or my dog snoring under the desk. Measured with a highly scientific iPhone 😊, the room, with nothing running is 45db and 60db with the A6000 running full fan.
@@HeatonResearch wow 60dB is pretty loud considering I saw 86 temperature.
Thanks for the info!
I wonder how that handles cyberpunk 2077?
It handles it pretty well, actually. instagram.com/p/CLAwFnyhN4p/
How does it compare to something like the 3090 or 3080? How does it do with frames per second?
Nice anti-static handling technique. I highly recommend it for all nvidia cards. Also nvidia cards may have a massive mis-match between compute and memory bandwidth for the newer and faster NN algorithms. For the NN segment of the market they face huge re-engineering costs. However I'm sure they have the money. Anyway it's not their fault. It's down to the broken foundations of NN research. The failure even to apply the scientific method to the dot product and other simple aspects of NNs.
Yeh, they don't even know how to use the dot product as a general associative memory by meeting the demands of the variance equation. Ie. Random projection, non-linear function, dot product. Terrible basic research. The NN books start in the middle and work forward. No one looked back to check the basics and challange assumptions.⚠⚠⚠
But can it run crysis ?
Isn't that always the question, though! :)
You beat me to it!
If you are wondering what the actual box looks like, if you get this card through more traditional retail, www.leadtek.com/eng/products/workstation_graphics(2)/NVIDIA_Quadro_RTX_A6000(30893)/detail
@Jeff Heaton, Hey man just a word of advice. Never daisy chain power connectors to your rtx cards. Yeah I know psu supply them but it's their way of cutting costs. Always use separate 8 pin connects for each plug in, if that makes sense. Yeah most won't care and do it daisy chaining, but if you're using it for high end rendering, ai simulations, neural networks, or in my case 4x 3090 FE's for mining, separate connections EACH to pardon my French, 'Cover your @$$'... See ya im outta here~
@@battosaijenkins946 Good advice
you should change the cooler to something else like noctua. that corsair aio isn't covering the whole cpu.
Nice.
You need a really good camera, the content it's amazing, you just need a bit of lightning and camera quality
Thanks, yes that is an evolving process.
Where did you order the PCIe to 8-pin CPU splitter cable?
It came with the GPU.
Quite a card ;-) Is it loud under stress?
Yes it is loud, have not specifically measured it yet..
look world gloves!! yup! 4 real !! nice presentation Jeff! popcorn engaged! the Angel on his shoulder says: hand with care!! the devil on the other shoulder, yells out. nail this card 2 the floor Jeff!! woohoo!! I'm like thinking, mastercard ''will'' cover & warranty this product!! though I ''think'' nVidia has a lifetime guarantee. but I'm not 4 sure!! good luck!!
👆 THEY GOT GPU ,3060,3070,3080,MANY MORE FOR A FAIR MARKET PRICE
What type of performance tradeoffs should I consider, if I use the RTX A6000 in an eGPU via Thunderbolt connection?
Can you please share the specs for the entire workstation? or if I can find it somewhere on your channel or blog?
Just updated the descriptions.
Awesome. I may go for broke and get one of these before the bots and scalpers pick them all up.
I feel your pain! :( 11:15
👆 GET FROM THEM ASAP GIVING OUT DISCOUNT ON ALL PRODUCT
I hope they release lower end such as a5000 and a4000, preferably with option for nvlink
It will be interesting to see how the axxxx series unfolds.
@@HeatonResearch nVidia has just released the A4000 & A5000 😃 The A4000 looks interesting with 16GB ram at only USD999, but has no nvLink 😥
btw how loud is the blower style cooler when running 100% for several minutes? I have heard they are much louder than regular cards?
Nice
How loud is the A6000 fan/blower? I’m keen to buy one but I’m not sure if it will distract my colleagues
Running at 80% it is about 60 decibels, as measured by my iPhone. By contrast, a silent room in my house is 40. In an office setting, you would hear it for sure, but not overly obnoxious (at least to me).
oh okay so this new A series is like their new brand of Quadro cards (or tesla) interesting
what are you using it for?
how is 3060 compare to google colab free version
It will be interesting to know what etherium mining hashrate on this card. Can you do the test?
Sure: th-cam.com/video/JyJiVWKApUA/w-d-xo.html
@@HeatonResearch I think every modern GPU should mine ETH now then not in use for other things. Especially then mining profit is over the roof now.
What is the difference in use cases and performance between this and an a100?
👆 THEY GOT GPU ,3060,3070,3080,MANY MORE FOR A FAIR MARKET PRICE
Hello. Is this GPU can be powered from ATX PSU directly with standard PCIE connector and without this Y-splitter cable? Thanks.
Yes, once I buy that cable.
Best X570 mother board for dual rtx A5000*2
(Cpu 5950x)
Would you recommend any list
👆 THEY ARE BEST HIT UP FOR A GOOD PRICE
Good for Plex?!
4 DISPLAY PORTS...not DVI
👆THEY GOT DOPE PC HIT UP
I'm insanely curious how many points per day Foldingathome would estimate on that card.
I never did get into that, but might be worth a try.
We're here for the seeing what the card can do. Why are you wasting 5 minutes just building the computer? There are other channels dedicated to pc builds.
👆👆 CHECK OUT FOR A FAIR PRICE
It is powerful but can it play cyberpunk2077¿
It does pretty well, actually... instagram.com/p/CLAwFnyhN4p/
Is this supposed to be better than RTX8000?
I've used both, and yes the A6000 is faster. Same memory. Right now the A6000 fills the same top spot in the professional workstation GPU line that the Quadro RTX 8000 filled. I am not sure if NVIDIA is holding the "A8000" name for an even more powerful card in this line.
I wonder how this compares to a 3090 or rtx 6000 in professional work loads.
I think that is a real question if you do not need the memory. I don't have the hardware to test that. I am looking forward to some 3090 benchmarks myself.
@@HeatonResearch great, I love your video keep up the great work👍
@@HeatonResearch Look above at my topmost reply ;)
Not many people can afford a A6000..so why dont you talk about the A4000 or the A5000.......which is more in line with what most people will use..
They did not exist at the time I made this video. I will probably cover them in a future video. I do have videos on cheaper cards, such as: th-cam.com/video/CPAmiRkP3U0/w-d-xo.html
@@HeatonResearch Thank you, would be good if you can do a video to supplement the latest RTX A0000 series, on the RTX A4000 and the RTX5000 seen as you've done one on the RTX6000. These are obviously the more affordable and latest Nvidia GPU's that most of us would be interested learning more about specifically and perhaps some demos on their performance...As the RTX A6000 is out of our budget....and most of us will never see the light of day actually using..... probably a good idea on the more affordable versions of its line. Thank you. Great video Jeff, keep up the great work
Actally 48 :)
No vr ready wtf
lol the bitlocker . just login to your microsoft account on another pc to get your bitlocker key .
i dont care what anyone says, blowers have always been the best for my system... such a pity this design has been pushed out. by far the best looking gpu... every other design has been hot garbage... too gamerish, screams im only 4 foot 5 but i have 22 inch tires with 8 wheel drive that can rock climb with tower like suspension cause it looks "kewl" with tons of RGB
Might wanna get some real fans in your case. A6000 with Corsair disco fans is not the look.
Good points, but so far so good with the system cooling, I do monitor GPU/CPU/RAM temps. The temperature is in an acceptable range and not throttling.
@@HeatonResearch Understood. Yet still a shame for such a beautiful and powerful GPU...to be disrespected with Corsair anything.
Nothing data science, just installation.
Gimme
👆GET FROM THEM FOR A FAIR PRICE