Jump to content

Increase your VRAM in 10 Steps with Regedit

I have experimented with this and it does work.

 

On my old laptop, an Asus Rog G571JT with a 970m, 24GB RAM and a i7 4710 (I think), I took RAM I didn't use and plugged it into a regedit key which boosted my VRAM performance.

 

You can do this with any Intel computer and at least 6GB+ of extra RAM.

 

1. Go to regedit

2. Local HKEY

3. Software

4. Intel

5. Create new Key as 'DVM'

6. Enter new Dword as 'dedicatedsegmentsize'

7. Set value in MB to above 6GB (1024x8=8192, 1024x10=10240, etc)

8. Click okay

9. Restart computer

10. Run bench to test visually

11. Visual performance Increase? Check.

 

Simply change the value to 0 and delete then restart again to revert.

 

EDIT here is a short test with Not using DVM vs using DVM. Note it is a short 8K test just to see some differences. Make sure to watch in 1080p.

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, Biggerboot said:

No experience with doing this but in theory I think the extra vRAM would be accessed more slowly.  Not that there isn't merit to it.

Load times and graphics performance actually Increase. It did for me.

Link to comment
Share on other sites

Link to post
Share on other sites

using system memory as vram? GREAT IDEA!!!!!!!!

 

disclaimer: that was sarcasm. system memory is very slow compared to vram. 

She/Her

Link to comment
Share on other sites

Link to post
Share on other sites

Then why did it boost my Performance?

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, aandril said:

Load times and graphics performance actually Increase. It did for me.

Maybe you would if it's faster than the last 512mb on the GTX 970 (I don't know if the 970m has the same quirk).

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, Biggerboot said:

Maybe you would if it's faster than the last 512mb on the GTX 970 (I don't know if the 970m has the same quirk).

I don't know what you mean? GTX 970m has 3GB VRAM.

Link to comment
Share on other sites

Link to post
Share on other sites

6 minutes ago, Biggerboot said:

Maybe you would if it's faster than the last 512mb on the GTX 970 (I don't know if the 970m has the same quirk).

No, they used a 192bit controller, not a 224+32 setup like the GTX 970. Meaning it just had one memory controller sliced off, hence the 3GB.

6 minutes ago, aandril said:

Then why did it boost my Performance?

Your test methodology could be wrong, it could be within variance. There's no reason for slow system memory (22-30gb/s) to add anything meaningful to the performance of a chip running >120gb/s memory. 

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, Majestic said:

No, they used a 192bit controller, not a 224+32 setup like the GTX 970. Meaning it just had one memory controller sliced off, hence the 3GB.

Your test methodology could be wrong, it could be within variance. There's no reason for slow system memory (22-30gb/s) to add anything meaningful to the performance of a chip running >120gb/s memory. 

I've tested this on both my 2014 Asus Rog and 2017 Alienware R4 17, and both have a boost in visual performance....

Link to comment
Share on other sites

Link to post
Share on other sites

The boost in performance can come from reducing the load on Windows memory management.  Less page shuffling for dynamic VRAM spillover because it is allocated all the time from boot time.

Link to comment
Share on other sites

Link to post
Share on other sites

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, Majestic said:

Explain visual performance?

Smoother performance (increased FPS, no FPS cuts), as well as faster loading.

For example on my Asus it was hard to run Unigin Superposition on high settings, but doing this helped.

Link to comment
Share on other sites

Link to post
Share on other sites

On 6/7/2018 at 1:11 PM, Biggerboot said:

Which games/benchmarks?

Look at my previous reply

Ah yes, and it also made graphics quality better as well.

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, aandril said:

Look at my previous reply

I'm not seeing any games or benchmarks listed.

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, aandril said:

Look at my previous reply

I see you added that in.

I would try some real games, because I think a synthetic benchmark may only take advantage of all the vRAM you have allocated and scale to that.

Link to comment
Share on other sites

Link to post
Share on other sites

On 6/7/2018 at 1:13 PM, KarathKasun said:

I'm not seeing any games or benchmarks listed.

My reply to Majestic.

I said Unigine Superposition on high settings with my Asus, for example.

 

On 6/7/2018 at 1:14 PM, Biggerboot said:

I see you added that in.
I would try some real games,

I've done so already. Same result. Better performance and visuals.

Link to comment
Share on other sites

Link to post
Share on other sites

4 minutes ago, aandril said:

My reply to Majestic.

 

I said Unigine Superposition on high settings with my Asus, for example.

Are we talking about a 5% higher score or a 1% higher score?  Or is it just more consistent, IE higher 0.1/0.01% FPS lows (less stutter)?

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, KarathKasun said:

Are we talking about a 5% higher score or a 1% higher score?  Or is it just more consistent, IE higher 0.1/0.01% FPS lows?

Higher score, better frames, better quality graphics....

Link to comment
Share on other sites

Link to post
Share on other sites

6 minutes ago, aandril said:

I've done so already. Same result. Better performance and visuals.

It'd be cool to share your results.  It's a neat little discovery but we'd like to see more.

Otherwise I'll experiment with it myself. 

Link to comment
Share on other sites

Link to post
Share on other sites

On 6/7/2018 at 1:21 PM, Biggerboot said:

It'd be cool to shared your results.  It's a neat little discovery but we'd like to see more.
Otherwise I'll experiment with it myself. 

I was thinking about posting a video on YouTube of the differences, but I was lazy. 

You can do it yourself yep. And pay attention to the game/bench itself first, rather than the numbers. So you can actually see the difference in quality.

 

Your temp should go down about 2 degrees also, if I remember correctly.

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, aandril said:

Your temp should go down about 2 degrees also, if I remember correctly.

If the dedicated vRAM is being less utilized in response to the change I could see that, but two degrees is within margin of error.

Link to comment
Share on other sites

Link to post
Share on other sites

All I'm saying is that it worked for me. If it works for you too, then great.

One extra thing to use for a performance boost.

Link to comment
Share on other sites

Link to post
Share on other sites

Did anyone happen to notice the path was "HKEY_LOCAL_MACHINE\Software\Intel"?

 

I'm pretty sure this only affects the Intel graphics memory, not any dedicated GPU memory. It's likely OP saw an apparent increase in performance because they're using a laptop which is likely using Optimus, and Optimus based laptops still output over the iGPU, not dGPU. So it was probably just a wee bit too low for good enough performance.

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×