How Much RAM Do Gamers Need? 8GB vs. 16GB vs. 32GB

[Nightfire]

It looks like the 8gb GTX580 is still the best choice for mid budget gamers as you could still get away with just 8gb of vram. This is not the case for 4gb cards and possibly not the 6 GB GTX 1060.

The higher end nVidia cards really demand alot of system ram. I know that the extra $200 to go from 16 GB to 32 GB is pocket change for someone getting a $1300 gpu, but for those spending around $700 for the gpu, that $200 is significant.

That 8s why the GTX 1080ti would be a much better buy than the GTX 2080 as the extra 3 GB of vram may make a difference for those running 'only' 16 GB of system ram.

 

[MonsterZero]

No reason to have 8GB anymore, that's not even enough for daily tasks for a business user.

With 3rd party bg apps for lighting, headphones, overclocking, synapse, CAM, etc I'd suggest 32GB, unless you want to hit above 75% of ram utilization while playing games.

Clearly you have no clue to what you are talking about. Today on my work PC I had the following open on my dual monitors all at once;

chrome with ten tabs open
firefox with ten tabs open
chrome with a single tab open
file explorer open
bitdefender up and running
notepad open
three separate versions of MS excel open each with their own spreadsheet
two PDFs

I have 6GB on my work PC and was only using 4.2GB of DDR3 RAM. I would say all those programs are 90% over what the average person has open on their laptop/PC but yeah 8GB is not enough for "daily tasks for a business user"....

Depends on your workload I guess but most business users are going to have:

Skype for Business, or some chat software
Excel, word, ppt or all of them
Microsoft outlook
onedrive, or dropbox
chrome, firefox or IE or all of them
Antivirus
3rd party client applications (which take up a majority of your memory)
Not to mention background services.

Do you even work bro?

 

[Nightfire]

No reason to have 8GB anymore, that's not even enough for daily tasks for a business user.

With 3rd party bg apps for lighting, headphones, overclocking, synapse, CAM, etc I'd suggest 32GB, unless you want to hit above 75% of ram utilization while playing games.

Clearly you have no clue to what you are talking about. Today on my work PC I had the following open on my dual monitors all at once;

chrome with ten tabs open
firefox with ten tabs open
chrome with a single tab open
file explorer open
bitdefender up and running
notepad open
three separate versions of MS excel open each with their own spreadsheet
two PDFs

I have 6GB on my work PC and was only using 4.2GB of DDR3 RAM. I would say all those programs are 90% over what the average person has open on their laptop/PC but yeah 8GB is not enough for "daily tasks for a business user"....

It seems the YouTube crowd has migrated over and spitting their non sense. Even today, 8 GB is still enough for most businesses despite what the peanut gallery says.

 

[Nightfire]

Decent article, although I do not like your choice of processor. I would think a more mainstream proc would have lead to a better standard. Not many people are going to be running that i9 proc. Decent nonetheless.

You are a Packer backer so I won't be too harsh on you. Calling this article only descent undermines how much effort goes into this type of testing. Going as far as resetting the system after every test and only using the last 60 seconds shows the level of attention to detail in order to get the most accurate results

You see, when testing components, it is really best to try and isolate one item. Here, Steve wanted to just test system memory usage in the newest games.
Using the 9900k eliminated any possible bottleneck there.

Even when testing the lower end cards, there is no reason to use a lesser CPU as it could confuse the results. There are plenty of CPU and game test reviews to give users knowledge on what kind of CPU is needed.

 

[torchyear]

No reason to have 8GB anymore, that's not even enough for daily tasks for a business user.

I work 2 separate office jobs and use 2 machines each running Win7 32-bit with 4GB RAM. 6+ Office apps / multi-tab browser / SQL database client, it's fine. It's the spinning hard drive transfer rate that's holding things back, rather than RAM. Taskman never maxes out.

 

[Adi6293]

I got 32GB od DDR4 3200Mhz plus Ryzen 7 2700X in my gaming rig, lots of ram and lots of threads, hopefully I don't have to upgrade for a while

 

[dirtyferret]

Depends on your workload I guess but most business users are going to have:

Skype for Business, or some chat software
Excel, word, ppt or all of them
Microsoft outlook
onedrive, or dropbox
chrome, firefox or IE or all of them
Antivirus
3rd party client applications (which take up a majority of your memory)
Not to mention background services.

Do you even work bro?

1. I'm not your "bro"
2. I just explained a higher workload then your potential workload and what that exact workload entails in RAM

 

[tipstir]

32GB vs 8GB and turn off the VRT But the system is always unstable.

 

[kmo911]

Now im at a Gigabyte Z390 GAMING SLI, S-1151 i7-9700

https://www.komplett.no/product/110...g-sli-s-1151?noredirect=true?noredirect=true#

https://www.komplett.no/search?q=i7-9700&nlevel=10000§28003§11204&fromSearchOverlay=1

WATER cooler o.c

gtx 1080 ti FE 11 gb version

only at 16 gb 2400 mhz DOH

 

[gamerk2]

I think you meant Deus Ex: Mankind Divided.

You're right. My b.

 

[transerv]

There Already working om DDR 5 for the Next Chips for Later Next Years 4 is Really Old so a Newer 5 will be needed also AL New M.B. with those chips. They say these NEW Chips and Ram WILL REALY BE FAST. Making the 4 a Dinosaur Like the 800 Ram, $$$$$$$

 

[cliffordcooley]

There Already working om DDR 5 for the Next Chips

I skipped DDR2 and likely going to do the same for DDR4.

 

[BrianMontanye]

You are a Packer backer so I won't be too harsh on you. Calling this article only descent undermines how much effort goes into this type of testing. Going as far as resetting the system after every test and only using the last 60 seconds shows the level of attention to detail in order to get the most accurate results

You see, when testing components, it is really best to try and isolate one item. Here, Steve wanted to just test system memory usage in the newest games.
Using the 9900k eliminated any possible bottleneck there.

Even when testing the lower end cards, there is no reason to use a lesser CPU as it could confuse the results. There are plenty of CPU and game test reviews to give users knowledge on what kind of CPU is needed.

Perfectly understand the bottleneck issue. The issue I have is this performance isn't indicative of what the average user will see when using those vid cards. That is all. I stand by my "decent" tag. I appreciate the work that is put in. But just decent as it more eye candy than usable info on the average Joe picking out a system and using this as helpful info. If you would have used a 7700k or even 8700k this would have been a much more defining example. Meaning, more people could actually relate to what is being presented. I am looking at it one way, and my comment was not a bash on the writers, just some useful feedback.

 

[moral hazard]

As a 12GB gamer I'm not sure what to make of this article. I also question their methodology. Using a worst case to predict what will happen at lower memory levels does not take into account what memory optimization and management steps have been implemented by the app developer, driver,and Windows when the actual situation arises. Seeing how much memory is used with 32GB ram / 8GB gram doesn't necessarily correlate with what is used at lower levels, much less with what you experience.

 

[imdarkbreeze]

" If you’re spending over $1000 on a graphics card, are you honestly going to think twice about spending $400 of DDR4 memory? Probably not."-quote


You're kidding right? EVERY single day we see people with a 400 dollar CPU, 800-1000 dollar graphics card, 300 dollars or more worth of RGB accessories and a 100 dollar case that want to use a 30 dollar power supply, so yes, these same people which are generally the BULK of the user base, ARE going to try and get away with using only 80 dollars worth of memory and a 60 dollar motherboard if they can.

In fact, not only that, they'll try using mixed sticks of the lowest quality, from their last build or wherever they can scrounge it up from for that matter.

 

[Steve]

" If you’re spending over $1000 on a graphics card, are you honestly going to think twice about spending $400 of DDR4 memory? Probably not."-quote


You're kidding right? EVERY single day we see people with a 400 dollar CPU, 800-1000 dollar graphics card, 300 dollars or more worth of RGB accessories and a 100 dollar case that want to use a 30 dollar power supply, so yes, these same people which are generally the BULK of the user base, ARE going to try and get away with using only 80 dollars worth of memory and a 60 dollar motherboard if they can.

In fact, not only that, they'll try using mixed sticks of the lowest quality, from their last build or wherever they can scrounge it up from for that matter.

I've not seen any evidence of this, it has to be the exception rather than the rule. Please point me in the direction of someone who bought an RTX 2080 Ti and has 8GB of RAM or less.

As a 12GB gamer I'm not sure what to make of this article. I also question their methodology. Using a worst case to predict what will happen at lower memory levels does not take into account what memory optimization and management steps have been implemented by the app developer, driver,and Windows when the actual situation arises. Seeing how much memory is used with 32GB ram / 8GB gram doesn't necessarily correlate with what is used at lower levels, much less with what you experience.

If you have 12GB's of DDR4 you're doing it wrong and I question your methodology

You say you're criticising the test method but really you're not. You are criticising how we looked at RAM allocation, not the actual benchmarking. I really should have spent some time explaining how the Windows pagefile works as well.

1. I'm not your "bro"
2. I just explained a higher workload then your potential workload and what that exact workload entails in RAM

Thanks Bro!

 

[kcwilsonii]

No reason to have 8GB anymore, that's not even enough for daily tasks for a business user.

With 3rd party bg apps for lighting, headphones, overclocking, synapse, CAM, etc I'd suggest 32GB, unless you want to hit above 75% of ram utilization while playing games.

Clearly you have no clue to what you are talking about. Today on my work PC I had the following open on my dual monitors all at once;

chrome with ten tabs open
firefox with ten tabs open
chrome with a single tab open
file explorer open
bitdefender up and running
notepad open
three separate versions of MS excel open each with their own spreadsheet
two PDFs

I have 6GB on my work PC and was only using 4.2GB of DDR3 RAM. I would say all those programs are 90% over what the average person has open on their laptop/PC but yeah 8GB is not enough for "daily tasks for a business user"....

try adding another 8gb world of difference. and for Microsoft office Applications use the 64 bit version not the 32 bit one

 

[kcwilsonii]

..Until then, my i5 4690K, Nvidia 970 and 8GB of 2400Mhz DDR3 will suffice, As it still does everything I need it to.

My PC is now about 6 years old and still runs all games, I just put in a 1060 last year. Running a i5 3570K, a 256gb SSD and 16gb of ram.

I'll probably upgrade once the nvidia 30x0 is out.

I have the same setup got 1060 off of ebay this year just to play a game. I may be looking at Intel in 2020 for a new buy on a video card. Nvidia has got to get those prices down

 

[cliffordcooley]

try adding another 8gb world of difference

If your system is not paging to slow media, there will be very little difference. If you really need to know if you need more memory turn your page file off. If you can run your system fine without a page file. Then you do not need more memory. It is a simple test.

I have the same setup got 1060 off of ebay this year just to play a game. I may be looking at Intel in 2020 for a new buy on a video card. Nvidia has got to get those prices down

And you think Intel will be cheaper or at this point better performance? I doubt either of those points.

 

[dirtyferret]

try adding another 8gb world of difference. and for Microsoft office Applications use the 64 bit version not the 32 bit one

drugs?? it would explain that answer

 

[dirtyferret]

Thanks Bro!

You can be my bro Steve, in a platonic way not that there's anything wrong with the other way.

 

[Steve]

You can be my bro Steve, in a platonic way not that there's anything wrong with the other way.

 

[dirtyferret]

don't get your hopes up, my wife says I can't date anymore

 

[cliffordcooley]

don't get your hopes up, my wife says I can't date anymore

No more bro nights?

 

[Nightfire]

Even though 32 GB showed a small improvement in some cases with the 2080ti, it does not make alot of sense since 8 GB was often enough for the Vega 64 in those same scenarios. I am thinking driver issues.

It would be interesting to see if quad-pumped memory would help the 2080ti. Unfortunately, you would need to use a slower CPU than the 9900k for that testing.

You will never satisfy our unquenchable thirst for benches Steve, but I think a lot more would like the focus to be on the midrange. As you have stated, the big boys are going to have 32gb anyhow, so what us the point of comparing less?

Going forward, I think there would be more use comparing 8gb, 12gb, and 16 GB systems when using mid level cards. It could help people really build the most cost efficient budget machine before hitting a ram bottleneck: an Rx580 (8gb) with 8gb system memory, a Gtx 1060 (6gb) with 12 gb of system memory or a cheaper RX480(4gb) with 16 GB of memory.