• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Should the type of memory be printed on the card box

Meh, doesn't bother me, hasn't done in the past and can't say that knowing now would sway me, as it really wouldn't. I don't overclock my cards anyway for real usage, nor do I bench really.

Give me a coil whine free card any day of the week, over a card with a certain make of VRAM.
 
Last edited:
I did not test memory on R9 290/X cards but I did some benches with different memory on 970 cards.May be of interest to some people.

1.970GTX G1 Gigabyte with Hynix Memory


2.970GTX G1 Gigabyte with Samsung Memory


1
Hynix
1404 Core
7008 Memory



Hynix
1537 Core
7600 Memory



Hynix
1593 Core
7600 Memory



Hynix Theif Benchmark
1593 Core
7600 Memory

2
Samsung
1404 Core
7008 Memory



Samsung
1537 Core
7600 Memory



Samsung
1593 Core
7600 Memory



Samsung Theif Benchmark
1593 Core
7600 Memory
 
Geeman1979 said:
Don't see why not. Clocks I just ran we're 1100/1350 which should be achievable by most and eliminates potential for any throttling going on.
Click to expand...

Please run stock too, 1250Mhz as this is the speed most will use. I will run some at this speed too when i get a spare moment.
 
Geeman1979 said:
Could do when I get time but shouldn't make any difference. Will also mean under clocking one of the cards memory which may hamper the results.
Click to expand...

The memorys different timings may affect things though, even more so when overclocked - and you rule out any chance of memory error correction. Coupled with the fact that most 290 cards operate at 1250Mhz.

This is what i should have done when i first tested rather than a overclocked to the limits test.
 
For high end benching cards (lightning/classified) yes, otherwise no, to the average end user it makes little difference.
 
The timings are going to be different any how, hence the difference in performance no doubt.

Anyhow...

Ran twice for each card to rule out margin of error.
290X's both Stock custom cooled so no throttling going on.

Cpu 3930k @ 4.4
Cards at 1100/1350 tested individually.
Ignore what 3dmark reads as it doesn't read the actual clocks correct.

Samsung.
http://www.3dmark.com/3dm/6053361

Elpida.
http://www.3dmark.com/3dm/6053527

So as we can see from Firestrike the Gpu scores are not that far off. Below a couple of 100 points difference.
 
AMDMatt said:
No i did not, but cool.

It would be nice to see a Hynix comparison in the mix as well so maybe you can Kaap can sync up to test a few games. Tomb Raider perhaps?
Click to expand...

It is not easy for me to test the Elpida cards on their own where they are as I can not get the PC to boot off slot 3 with the others turned off.

Worse still when I tried it took out my bios setup and it has taken me more than an hour to get it running again.

An easy way to compare though is 4 up on the benching threads where even having 2 Elpida cards in the mix again your all Hynix cards gives a drop in performance at the same clocks.
 
AMDMatt said:
I saw a 1,000 point difference between Elpida at 1625Mhz and Hynix at 1625Mhz.

Unfortunately the testing was flawed as the Elpida could have had memory error correction kicking in and I never tested at stock.

I'm sure Kaap can test this though and confirm if there's any truth to it or not at stock as he has Hynix and Elpida memory on his 290X's.

I would expect to see a small difference (100-200 points) with both cards at 1250Mhz.

Kaap, over to you. :)

EDIT

PG Tips could test this too, but vs Samsung. :)
Click to expand...

Geeman1979 said:
The timings are going to be different any how, hence the difference in performance no doubt.

Anyhow...

Ran twice for each card to rule out margin of error.
290X's both Stock custom cooled so no throttling going on.

Cpu 3930k @ 4.4
Cards at 1100/1350 tested individually.
Ignore what 3dmark reads as it doesn't read the actual clocks correct.

Samsung.
http://www.3dmark.com/3dm/6053361

Elpida.
http://www.3dmark.com/3dm/6053527

So as we can see from Firestrike the Gpu scores are not that far off. Below a couple of 100 points difference.
Click to expand...

Interesting Gee.

Also, called it. (sort of) :cool: :D
 
Geeman1979 said:
The timings are going to be different any how, hence the difference in performance no doubt.

Anyhow...

Ran twice for each card to rule out margin of error.
290X's both Stock custom cooled so no throttling going on.

Cpu 3930k @ 4.4
Cards at 1100/1350 tested individually.
Ignore what 3dmark reads as it doesn't read the actual clocks correct.

Samsung.
http://www.3dmark.com/3dm/6053361

Elpida.
http://www.3dmark.com/3dm/6053527

So as we can see from Firestrike the Gpu scores are not that far off. Below a couple of 100 points difference.
Click to expand...

You mean 172 points on the graphics score.:)
 
I'm not sure if it would be possible to print on the box as some brands change Memory type depending is what is available at the time, and the posibility of mismatched box's and cards would be huge imo.
Wasnt it one of the classifieds even that changed for awhile simply because what was usual wasnt available
 
triss said:
I'm not sure if it would be possible to print on the box as some brands change Memory type depending is what is available at the time, and the posibility of mismatched box's and cards would be huge imo.
Wasnt it one of the classifieds even that changed for awhile simply because what was usual wasnt available
Click to expand...

Nothing to stop them putting a sticker on the box then.:)
 
You must log in or register to reply here.
Back
Top Bottom