Welcome Guest ( Log In | Register )

Outline · [ Standard ] · Linear+

 GPU Discussion V2

views
     
Acid_RuleZz
post Sep 27 2013, 04:17 PM

ミウ ❤
*******
Senior Member
6,612 posts

Joined: Jan 2003
From: Tomorrow


QUOTE(Kendall @ Sep 27 2013, 04:12 PM)
Because I don't want to spent too much money ma. I want the acer because it will come with a stylus for my study sad.gif I'm poor  cry.gif
*
Can you do a favor for me? Download 3DMark11, run the benchmark and post the score here.
Acid_RuleZz
post Mar 13 2014, 11:38 PM

ミウ ❤
*******
Senior Member
6,612 posts

Joined: Jan 2003
From: Tomorrow


QUOTE(holloway @ Mar 13 2014, 11:31 PM)
nice article. although i'm abit concerned as why they just rebrand 880m unlike 860m which went for maxwell tech. probably they are going going to skip maxwell and opt for broadwell in coming months for release of 980?
You meant Volta? Broadwell is Intel microarchitecture.
Acid_RuleZz
post Mar 17 2014, 09:18 PM

ミウ ❤
*******
Senior Member
6,612 posts

Joined: Jan 2003
From: Tomorrow


QUOTE(Mr_47 @ Mar 17 2014, 08:42 PM)
im confuse... sure 860m can beat 755m sli?

wtf lasttime saw kennot beat one. sifu helpppp!

http://www.notebookcheck.net/Mobile-Graphi...List.844.0.html
*
755m sli is faster if it scale perfectly. hmm.gif
Acid_RuleZz
post Mar 18 2014, 01:03 PM

ミウ ❤
*******
Senior Member
6,612 posts

Joined: Jan 2003
From: Tomorrow


QUOTE(maxzhengab @ Mar 18 2014, 12:14 PM)
any idea that gtx 660 is still good currently?wanted to buy a good graphic card..RM600 budget
*
This is mobile graphic section, head to hardware section. hmm.gif

GTX660 for RM600 is good buy imo, but you better wait for Maxwell successor for GTX660 or just grab GTX750 Ti.

QUOTE(holloway @ Mar 18 2014, 12:25 PM)
probably that 860m isn't at it's mature state? coz still new.. newer drivers with better support and optimizing are coming out. probably 860m single card is better  rclxms.gif Core is already slightly higher. but maybe for higher resolution 755m sli is better?

just wondering what do u mean scale perfectly.
*
Maybe, 860M is literally a GTX750 Ti so we know what to expect, hopefully thermal performance is good.

Most of the time multi-gpu don't scale 2x the performance. Usually you only get 1.5x-1.8x of performance or worst case you'll get -ve scaling where single 755m is better than 755m Sli.
Acid_RuleZz
post Apr 28 2014, 08:05 PM

ミウ ❤
*******
Senior Member
6,612 posts

Joined: Jan 2003
From: Tomorrow


QUOTE(herpmerpderp @ Apr 28 2014, 06:56 PM)
The current "Maxwell" GPUs are still based on Kepler architecture (28nm). The real Maxwell (20nm) has no ETA yet
*
What you mean by current Maxwell still based on Kepler?

There's no 20nm GPU this year, so for those that waiting... better stop waiting. laugh.gif
Acid_RuleZz
post Apr 29 2014, 11:03 AM

ミウ ❤
*******
Senior Member
6,612 posts

Joined: Jan 2003
From: Tomorrow


QUOTE(herpmerpderp @ Apr 29 2014, 02:38 AM)
I mean that the current Maxwell, is based on Kepler's architecture, not the proposed 20nm. Not even, the enthusiast cards are higher tweaked rebrands of the previous gen, with the slight exception of the 860m. Slight because there are Kepler variants of it being circulated. Was just a response to a posted a few posts up from mine.

And yes raymond, Moore's Law can only stretch so far.
*
Doesn't matter if it's on 28nm or 20nm, it still a different architecture. Maxwell on 28nm doesn't offer much performance increase but it does give much better performance per tdp compare to Kepler.

QUOTE(raymondtong94 @ Apr 29 2014, 09:58 AM)
i think i know what you mean. what you are trying to say is, current generation of potentially Maxwell cards turns out to be Kepler again.
actually, it is a tactic in business. its more like they want to clear out old stocks of the chip, so we cant do anything about it. as a customer, either option is go for AMD or wait.
*
No, both AMD and Nvidia can't jump to 20nm process node because TSMC hit few walls that make them have to delay it this year.
Acid_RuleZz
post Apr 29 2014, 03:58 PM

ミウ ❤
*******
Senior Member
6,612 posts

Joined: Jan 2003
From: Tomorrow


QUOTE(herpmerpderp @ Apr 29 2014, 03:52 PM)
Based on. And also the getting rid of surplus GPU's still ring true. Like I said, a fair portion of them are tweaked rebrands. Like the 870M is a tweaked 780m, or heavily based off it. We might see less of them as the year goes on. As for the 20nm chips, some say that even in 2015, to expect it much later in the year  sad.gif
Doubt I'll be plonking down the cash for a good laptop this year.

Recently cancelled my pre-orders on the Blade 14 for the same reasons. Even with all that R&D, I don't know how it will effectively cool an 870M without having to throttle. May as well buy a half decent laptop and save, wait till next year and splurge on a maxed out laptop worth the money.
*
But Pascal will arrive in 2016. tongue.gif

Nvlink, stacked DRAM, Unified memory. Much greater feature compare to Maxwell. lust.gif
Acid_RuleZz
post Jul 7 2016, 05:01 PM

ミウ ❤
*******
Senior Member
6,612 posts

Joined: Jan 2003
From: Tomorrow


QUOTE(CookiePirate @ Jul 7 2016, 02:26 PM)
What's the difference between 940m and 940mx eh?
*
Slightly faster clock speed compare to 940M (Maxwell).

Better chip architecture compare to 940M (Kepler).
Acid_RuleZz
post Dec 24 2017, 07:59 PM

ミウ ❤
*******
Senior Member
6,612 posts

Joined: Jan 2003
From: Tomorrow


QUOTE(tyan1234 @ Dec 24 2017, 04:49 PM)
Do you guys have any idea is Geforce Nvidia 1040 better or Nvidia Geforce mx 150 better? I couldn’t find any official data regarding Nvidia Geforce 1040 yet. However it was released by Lenovo in 320s model
*
Let see..

MX150 = 1030

1040>1030

1040 Win.

Win by how much? We don't know yet.

7 Pages « < 5 6 7Top
 

Change to:
| Lo-Fi Version
0.0360sec    0.78    7 queries    GZIP Disabled
Time is now: 12th December 2025 - 07:34 PM