• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

MSI GeForce RTX 3090 Suprim X

Joined
Apr 16, 2019
Messages
632 (0.34/day)
oh, so if nvidia does it, then power consumption doesn't matter? got it, thanks!
Not really for 1.5 - 2k $ cards, no; 295x2 for instance approached 700W under stress test

I wouldn't be so sure about that. The 8nm Samsung and 7nm TSMC are fairly similar in density. There are pros and cons on both sides and NV went for better yields and lower price per wafer instead of slightly better power efficiency. Either way it would have been better for NV to go TSMC but saying absolutely no competition if they would have gone 7nm is over exaggerated.
I'd be pretty sure; there already isn't that much competition when looking at the whole picture - particularly the horrendously bad ray tracing performance of RDNA2, which will definitely become much more important before the life cycle of current cards is over (I'd wager that in two years or less more than half of new titles will implement some sort of RT functionality) and last but definitely not least, DLSS which will definitely help to prolong the usability of Ampere cards especially in 4k, but also lower resolutions further down the line. Now if we imagine considerably better power efficiency coupled with notably higher clocks on top of this (from 7nm TSMC), I don't think what I said is over exaggerated at all (although obviously, AMD would have to position their MSRPs much lower in that case)
 
Last edited:
Joined
Jul 9, 2015
Messages
3,413 (1.05/day)
System Name M3401 notebook
Processor 5600H
Motherboard NA
Memory 16GB
Video Card(s) 3050
Storage 500GB SSD
Display(s) 14" OLED screen of the laptop
Software Windows 10
Benchmark Scores 3050 scores good 15-20% lower than average, despite ASUS's claims that it has uber cooling.
The 8nm Samsung and 7nm TSMC are fairly similar in density.
I keep hearing that 8nm Samsung is 10nm rebrand, so, shrug.

Except that that this card is a HUGE jump in performance over the RTX2000 series cards, something that didn't happen with the Fermi gen cards
An interesting myth.

No, performance was not the issue with Fermi, power consumption and price was. Both are there with Ampere.
 
Joined
May 31, 2016
Messages
4,338 (1.49/day)
Location
Currently Norway
System Name Bro2
Processor Ryzen 5800X
Motherboard Gigabyte X570 Aorus Elite
Cooling Corsair h115i pro rgb
Memory 16GB G.Skill Flare X 3200 CL14 @3800Mhz CL16
Video Card(s) Powercolor 6900 XT Red Devil 1.1v@2400Mhz
Storage M.2 Samsung 970 Evo Plus 500MB/ Samsung 860 Evo 1TB
Display(s) LG 27UD69 UHD / LG 27GN950
Case Fractal Design G
Audio Device(s) Realtec 5.1
Power Supply Seasonic 750W GOLD
Mouse Logitech G402
Keyboard Logitech slim
Software Windows 10 64 bit
I'd be pretty sure; there already isn't that much competition when looking at the whole picture - particularly the horrendously bad ray tracing performance of RDNA2, which will definitely become much more important before the life cycle of current cards is over (I'd wager that in two years or less more than half of new titles will implement some sort of RT functionality) and last but definitely not least, DLSS which will definitely help to prolong the usability of Ampere cards especially in 4k, but also lower resolutions further down the line. Now if we imagine considerably better power efficiency coupled with notably higher clocks on top of this (from 7nm TSMC), I don't think what I said is over exaggerated at all (although obviously, AMD would have to position their MSRPs much lower in that case)

I don't think it's bad but it does lack in comparison to NV's cards. I also don't think that Ray Tracing matters the most here at what the state of Ray Tracing is at this point in terms of support for software. But that's just how I see it at this point in time. That is your wager but we can't be certain. Even though the NV RT performance is better than AMD's, it isn't is great for certain games. So if your wager is correct and there will be plenty of new games with RT my wager is they would require more RT performance which current cards either NV or AMD may not have (that includes DLSS like feature)
The thing is, TSMC and NV Ampere graphics card, You don't know how much of an improvement will be there to be honest. It may have been same performance as Samsung's but less power draw.

BTW. The R9 295x2 was dual Graphics chip card. Don't think that is relevant and on a node from the stone age. So if you compare 3090 over 500W peak, the R9 295x2 doesn't look bad as for a dual card and chip done in 28nm.
I keep hearing that 8nm Samsung is 10nm rebrand, so, shrug.
Well yes that's what people say but you already know that 7nm or 8nm is just a naming scheme nothing more.
 
Last edited:
Joined
Jul 5, 2013
Messages
25,690 (6.45/day)
An interesting myth.
No, performance was not the issue with Fermi, power consumption and price was.
Oh really? Let's review.
The first shows Crysis with a GTX260, the next a GTX560. Pay attention to the 1920x1200 scores in particular.

Now let's look at power usage, shall we?
Isn't that interesting?

Now lets look at the top tier models for each of those generations;
Hmm..

Now look at the power consumption for each;
Well isn't that interesting also?

You were saying what now?
Context is important. Try it sometime. You too @ratirt

Ampere is NOT Fermi 2.0. Perfect rubbish notion. To make such a claim is as ridiculous as it is completely lacking in merit.
 
Joined
Jul 9, 2015
Messages
3,413 (1.05/day)
System Name M3401 notebook
Processor 5600H
Motherboard NA
Memory 16GB
Video Card(s) 3050
Storage 500GB SSD
Display(s) 14" OLED screen of the laptop
Software Windows 10
Benchmark Scores 3050 scores good 15-20% lower than average, despite ASUS's claims that it has uber cooling.
Hmm?



285 - 26.1FPS
580 - 47fFPS

80% more perf is not a huge perf jump, but 45% more perf is:



:peace:

By this metric, Ampere Fermi 2 is 2 times worse than normal Fermi, as it pushed perf by only half of the original.
 
Joined
Apr 1, 2017
Messages
420 (0.16/day)
System Name The Cum Blaster
Processor R9 5900x
Motherboard Gigabyte X470 Aorus Gaming 7 Wifi
Cooling Alphacool Eisbaer LT360
Memory 4x8GB Crucial Ballistix @ 3800C16
Video Card(s) 7900 XTX Nitro+
Storage Lots
Display(s) 4k60hz, 4k144hz
Case Obsidian 750D Airflow Edition
Power Supply EVGA SuperNOVA G3 750W
That's not what was said, don't twist words out of context. AMD's current line-up is just as power hungry with very little variation.

oh, so AMD's current line-up uses less power than ampere, but it's using just as much power as ampere?! interesting!
 
Joined
Apr 16, 2019
Messages
632 (0.34/day)
oh, so AMD's current line-up uses less power than ampere, but it's using just as much power as ampere?! interesting!
They are not far apart and even if you push 6800XT to high heaven where it consumes at least as much as 3080, it still doesn't match the latter's performance:
 
Joined
Apr 1, 2017
Messages
420 (0.16/day)
System Name The Cum Blaster
Processor R9 5900x
Motherboard Gigabyte X470 Aorus Gaming 7 Wifi
Cooling Alphacool Eisbaer LT360
Memory 4x8GB Crucial Ballistix @ 3800C16
Video Card(s) 7900 XTX Nitro+
Storage Lots
Display(s) 4k60hz, 4k144hz
Case Obsidian 750D Airflow Edition
Power Supply EVGA SuperNOVA G3 750W
They are not far apart and even if you push 6800XT to high heaven where it consumes at least as much as 3080, it still doesn't match the latter's performance:
don't care LOL one card is 500W the other isn't
 
Joined
Apr 12, 2013
Messages
6,814 (1.68/day)
Fanboy or not the same garbage is being recycled as fact. I just posted links to "fact based review" a while back, waiting for the perpetrators to present their evidence!
 
Joined
Apr 1, 2017
Messages
420 (0.16/day)
System Name The Cum Blaster
Processor R9 5900x
Motherboard Gigabyte X470 Aorus Gaming 7 Wifi
Cooling Alphacool Eisbaer LT360
Memory 4x8GB Crucial Ballistix @ 3800C16
Video Card(s) 7900 XTX Nitro+
Storage Lots
Display(s) 4k60hz, 4k144hz
Case Obsidian 750D Airflow Edition
Power Supply EVGA SuperNOVA G3 750W
I was actually considering getting a 3000 series nvidia GPU, but the power consumption turned me off entirely
 
Last edited by a moderator:
Joined
Jul 9, 2015
Messages
3,413 (1.05/day)
System Name M3401 notebook
Processor 5600H
Motherboard NA
Memory 16GB
Video Card(s) 3050
Storage 500GB SSD
Display(s) 14" OLED screen of the laptop
Software Windows 10
Benchmark Scores 3050 scores good 15-20% lower than average, despite ASUS's claims that it has uber cooling.
I keep hearing this BS from multiple forum members, on various forums! Evidence please, not conjecture or your best guesstimate :rolleyes:

We went from "AMD is so many years behind" to "AMD will maybe compete with 2080Ti, but I doubt it" to full throttle meltdowns.

Have mercy.
 
Joined
Jul 5, 2013
Messages
25,690 (6.45/day)
Hmm?



285 - 26.1FPS
580 - 47fFPS
Hmm, let's review;
No, performance was not the issue with Fermi, power consumption and price was.
Yup, that's what you said... Nice cherry pick though, in doing so you effectively proved the point I was making, on several levels. Well done!

oh, so AMD's current line-up uses less power than ampere, but it's using just as much power as ampere?! interesting!
Um, YOU are not paying attention!
Not really for 1.5 - 2k $ cards, no; 295x2 for instance approached 700W under stress test
Henry was referring to AMD's previous line up for comparison.
And that shows that when I said...
AMD's current line-up is just as power hungry with very little variation.
...I was spot-on. 328W VS 324W qualifies as very little variation.

Do you two want to continue trying and failing to pick nits?
 
Last edited:
Joined
Apr 1, 2017
Messages
420 (0.16/day)
System Name The Cum Blaster
Processor R9 5900x
Motherboard Gigabyte X470 Aorus Gaming 7 Wifi
Cooling Alphacool Eisbaer LT360
Memory 4x8GB Crucial Ballistix @ 3800C16
Video Card(s) 7900 XTX Nitro+
Storage Lots
Display(s) 4k60hz, 4k144hz
Case Obsidian 750D Airflow Edition
Power Supply EVGA SuperNOVA G3 750W
Hmm, let's review;

Yup, that's what you said... Nice cherry pick though, in doing so you effectively proved the point I was making, on several levels. Well done!


Um, YOU are not paying attention!

Henry was referring to AMD's previous line up for comparison.
And that shows that when I said...

...I was spot-on. 328W VS 324W qualifies as very little variation.

Do you two what to continue trying and failing to pick nits?

where did you find that 328W VS 324W figure LOL
in the review you linked peak is 324W vs 350W, and average is 279W vs 339W, you can guess which is which ;)
are you shilling on purpose or are you being ignorant in regards to what you post on purpose?
 
Low quality post by spnidel
Joined
Apr 1, 2017
Messages
420 (0.16/day)
System Name The Cum Blaster
Processor R9 5900x
Motherboard Gigabyte X470 Aorus Gaming 7 Wifi
Cooling Alphacool Eisbaer LT360
Memory 4x8GB Crucial Ballistix @ 3800C16
Video Card(s) 7900 XTX Nitro+
Storage Lots
Display(s) 4k60hz, 4k144hz
Case Obsidian 750D Airflow Edition
Power Supply EVGA SuperNOVA G3 750W
Joined
Jul 9, 2015
Messages
3,413 (1.05/day)
System Name M3401 notebook
Processor 5600H
Motherboard NA
Memory 16GB
Video Card(s) 3050
Storage 500GB SSD
Display(s) 14" OLED screen of the laptop
Software Windows 10
Benchmark Scores 3050 scores good 15-20% lower than average, despite ASUS's claims that it has uber cooling.
Joined
Jul 5, 2013
Messages
25,690 (6.45/day)
What the heck, dude:



what is it, on average 56% faster? Yet "performance was not there", but it's there with 45%... :D
You're still missing the context. READ;
No, performance was not the issue with Fermi, power consumption and price was. Both are there with Ampere.
Yup, those are your words. Would you like to continue contradicting yourself?
 
Joined
Dec 4, 2020
Messages
21 (0.02/day)
Location
Sino
Processor R7 5700G R5 3500X
Motherboard ASrock X570 Pro4 ASUS B550M ZAKU
Cooling Thermalright PA120 YAJUN B3
Memory CRUICAL
Video Card(s) IGPU Leadtek RTX3080 Life ES
Storage INTEL OPTANE 900P 280G/KST A2000 1T/SX8200PRO 1T
Display(s) LG 27GL830
Case SAMA
Audio Device(s) hyperX alpha/microlab M200
Power Supply PCCOOLER GI-P850 LIANLI CS700
Mouse logitech G102 LIGHTSYNC
Keyboard LENOVO KB1021עִבְרִית
Software Microsoft Windows 10 (10.0) Pro for Workstations 64-bit
Benchmark Scores https://valid.x86.fr/l9x6tj
well done MSI:roll:
 
Joined
Dec 4, 2020
Messages
21 (0.02/day)
Location
Sino
Processor R7 5700G R5 3500X
Motherboard ASrock X570 Pro4 ASUS B550M ZAKU
Cooling Thermalright PA120 YAJUN B3
Memory CRUICAL
Video Card(s) IGPU Leadtek RTX3080 Life ES
Storage INTEL OPTANE 900P 280G/KST A2000 1T/SX8200PRO 1T
Display(s) LG 27GL830
Case SAMA
Audio Device(s) hyperX alpha/microlab M200
Power Supply PCCOOLER GI-P850 LIANLI CS700
Mouse logitech G102 LIGHTSYNC
Keyboard LENOVO KB1021עִבְרִית
Software Microsoft Windows 10 (10.0) Pro for Workstations 64-bit
Benchmark Scores https://valid.x86.fr/l9x6tj
I prefer to see some more silent and efficiency GPU/CPU in this generation gaming rigs.And i will turn to buy a console,maybe?
(i guess intel and nvidia are responsible for this.
 
Joined
Sep 26, 2009
Messages
82 (0.02/day)
Location
North Central Ohio
System Name Custom build - I built it for gaming and some content creation
Processor Intel i7 12700K on air
Motherboard MSI MAG Z690 Tomahawk WiFi DDR4
Cooling Noctua NH-D15S
Memory 64GB G.Skill 3600 DDR4 [16GB x 4]
Video Card(s) MSI RTX 3070 Ti SUPRIM X, power limit 75% & Mem +300
Storage SSD: Sabrent Rocket 2TB, Samsung 970 EVO 1TB, SSD Samsung 860 EVO 1T
Display(s) 55" Samsung Q80T 120Hz at 4K [on my desk]
Case Cooler Master 922 HAF [3rd 200mm fan in door], 10 years old & fans still run
Audio Device(s) Creative Sound Blaster Z
Power Supply Seasonic Focus Platinum PX-850
Mouse Corsair M55 RGB Pro
Software Windows 10 Pro 64-bit 21H2
Benchmark Scores [GPU Score] Time Spy: 13 925 [Power Limit 75%, MEM +300, GPU Max Power 232w, Max Hot Spot 62.4C
80'c ? pretty bad. ASUS RTX 3090 STRIX OC = 68'c

ASUS RTX 3090 STRIX OC (Quiet BIOS) was 75°C

 
Joined
Nov 2, 2013
Messages
463 (0.12/day)
System Name Auriga
Processor Ryzen 7950X3D w/ aquacomputer cuplex kryos NEXT with VISION - acrylic/nickel
Motherboard Asus ProArt X670E-CREATOR WIFI + Intel X520-DA2 NIC
Cooling Alphacool Res/D5 Combo •• Corsair XR7 480mm + Black Ice Nemesis 360GTS radiators •• 7xNF-A12 chromax
Memory 96GB G.Skill Trident Z5 RGB @ 6200MHz, 30-40-40-28, 1.35V (F5-6400J3239F48GX2-TZ5RK)
Video Card(s) MSI RTX 4090 Suprim Liquid X w/ Bykski waterblock
Storage 2TB inland TD510 Gen5 ••• 4TB WD Black SN850X ••• 40TB UNRAID NAS
Display(s) Alienware AW3423DWF (3440x1440, 10-bit @ 145Hz)
Case Thermaltake Core P8
Power Supply Corsair AX1600i
Mouse Razer Viper V2 Pro (FPS games) + Logitech MX Master 2S (everything else)
Keyboard Keycult No2 rev 1 w/Amber Alps and TX stabilizers on a steel plate. DCS 9009 WYSE keycaps
Software W11 X64 Pro
Benchmark Scores https://valid.x86.fr/c3rxw7
It's not great, no, but just like with Fermi, they are the undeniably faster cards and in this price bracket that matters (far) more than a couple dozen W of power consumption. If Ampere was on the same 7nm TSMC node though, there would once again be absolutely no competition...however the already bad availability would likely reach epic levels...

if nvidia would have went with 7nm TSMC, the GPU supply shortage would be magnitudes worse. we'd be paying $1500 for a 3080/6800xt, assuming we'd even see any 6800xt at all. nvidia, in their quest to save some money, actually ended up doing us all a favor. i'm not going to pretend i'm an engineer (like some of you here) so i won't pretend to know how well 8nm samsung compares to 10nm samsung or TSMC, HOWEVER, it's clearly better when amd and nvidia don't have to compete over the same node, along with gaming consoles and other contracts TSMC has to fulfil.

oh, so if nvidia does it, then power consumption doesn't matter? got it, thanks!

are we judging all ampere cards based on the performance of one garbage brand that makes garbage products? asus and gigabyte cards pull 100W less than the "suprim" 3090

i personally have a ryzen 5900x, 3090 tuf OC, with a custom loop running a d5 pump, 8 fans, peripherals, plus the obligatory RGB puke, ect. the higher power draw registered on my UPS so far was 512W for the ENTIRE system.

the difference in power consumption between reference 3080 and 6800xt, as well as 3090 and 6900xt, is about 60W. but don't forget that ampere is still faster on average and it can do what navi can't. the extra hardware present on ampere allows for nvenc, faster RT, rtx voice, dlss, etc. big navi is just a faster navi, it doesn't do anything special aside from render frames faster.

now, if you think that the extra features are not worth the extra power draw, then you decision on which card to buy is clear. if i had a choice between a 3080 and a 6800xt, at their proper pricing, i'd pick the 3080 100 times out of 100. but during the fermi days, i had amd cards. i had a 5970 followed by 2x 6950s unlocked with the 6970 BIOS
 
Joined
Aug 8, 2017
Messages
14 (0.01/day)
Power usage debate makes no sense to me. As long as it’s not heating up your room, you are spending thousands on high end parts so power cost is too low comparatively to care about. Noise is comparable to other gens.
So unless you’re over clocking, you plug the thing in and get FPS. You would have to be a special-use person to care about power.

the only reason power should matter is if it affects your life somehow. Otherwise it’s hypothetical debates.
 
Joined
Nov 2, 2013
Messages
463 (0.12/day)
System Name Auriga
Processor Ryzen 7950X3D w/ aquacomputer cuplex kryos NEXT with VISION - acrylic/nickel
Motherboard Asus ProArt X670E-CREATOR WIFI + Intel X520-DA2 NIC
Cooling Alphacool Res/D5 Combo •• Corsair XR7 480mm + Black Ice Nemesis 360GTS radiators •• 7xNF-A12 chromax
Memory 96GB G.Skill Trident Z5 RGB @ 6200MHz, 30-40-40-28, 1.35V (F5-6400J3239F48GX2-TZ5RK)
Video Card(s) MSI RTX 4090 Suprim Liquid X w/ Bykski waterblock
Storage 2TB inland TD510 Gen5 ••• 4TB WD Black SN850X ••• 40TB UNRAID NAS
Display(s) Alienware AW3423DWF (3440x1440, 10-bit @ 145Hz)
Case Thermaltake Core P8
Power Supply Corsair AX1600i
Mouse Razer Viper V2 Pro (FPS games) + Logitech MX Master 2S (everything else)
Keyboard Keycult No2 rev 1 w/Amber Alps and TX stabilizers on a steel plate. DCS 9009 WYSE keycaps
Software W11 X64 Pro
Benchmark Scores https://valid.x86.fr/c3rxw7
Power usage debate makes no sense to me. As long as it’s not heating up your room, you are spending thousands on high end parts so power cost is too low comparatively to care about. Noise is comparable to other gens.
So unless you’re over clocking, you plug the thing in and get FPS. You would have to be a special-use person to care about power.

the only reason power should matter is if it affects your life somehow. Otherwise it’s hypothetical debates.
so, 2 things.

first, i upgraded from a 2080 ti to an msi 3090 trio and It actually was heating my room up noticeably. at no point before did i have to stop playing games because sitting next to my computer become uncomfortable. i live in the north east US and this was in the month of December. my computer literally became a space heater. so yes, it did affect my life. the suprim draw even more power than the trio. so i can't imagine it would be any better

and second, there is no reason for the same product (rtx 3090) from two different companies, that perform about the same, to have a difference of 100W in power draw between the two. the msi card is garbage. 3090 strix is slightly faster (about 1% according to TPU) than the 3090 suprim but pulls 100W less power.
 
Top