Wish list for TPU VGA reviews

[DanishDevil]

It might not be badly programmed. It cloaks everything under it, and cloaked units have a pretty cool effect on them. Maybe you could run a replay through a super crazy GPU rig and see what you think (you know better than anybody else here).

 

[blu3flannel]

I thought Metro 2033 was already in there.

My apologies then. I was going to suggest the most-played game on Steam, then I looked it up and discovered it was Counter-Strike. In a close second is CS:S. What about Civ V?

 

[W1zzard]

What about Civ V?

how about people look at my reviews before asking questions?

 

[cq842000]

Apologies if this has already been asked, but are benchmarks like Unigine Heaven out of the question for testing power consumption?

 

[SuperStarr]

Wizz, i know it is unoptimized but can you add World of Tanks in your benchmarks?

 

[gumpty]

Oh and while you're at it, perhaps get a 2560x1440? jk (I compare using your 2560x1600 just fine)

I want 2048 x 1152.

 

[LAN_deRf_HA]

If I had to guess I'd say 2560x1440 will be the next popular resolution. http://accessories.us.dell.com/sna/productdetail.aspx?c=us&l=en&s=dhs&cs=19&sku=224-8284&~ck=baynoteSearch&baynote_bnrank=14&baynote_irrank=0

 

[Wyverex]

W1z, could you add all CCC Overdrive limits (minimums in addition to maximums)?

 

[blu3flannel]

how about people look at my reviews before asking questions?

I meant for typical power consumption. According to Steam, the peak players in-game today in Crysis 2 (ranked #55) is 720. Civ V (ranked #4), on the other hand, has a peak player count of 15,482. Since far more people play Civ V than Crysis 2, that would make it a better benchmark for typical power consumption since it's a typical game that is more representative of the gaming population.

 

[crazyeyesreaper]

and yet Crysis will offer a more reproducable load and thus makes more sense, and while CIV 5 might be more played it dosent make it better, if we want to get into most played games we might as well use farmville then / argument,

Ill take W1zzards ideas and thoughts over a normal user as he has been reviewing GPUs for what 7 years now?

besides using Crysis 2 for power usage isnt all that bad its a typical load altho he could use Metro 2033 as it is the closest to Furmark power usage without tripping GPU protections that are in place. eitherway dosent matter to me as W1zzard is STILL the only review to actually give us GPU power usage only something no one else does.

 

[dj-electric]

Overclocking on unlocked HD6950 GPUS, overclocking them without unlocking is almost irrelevant.

 

[theJesus]

I'm also interested in seeing multi-monitor reviews for the super high-end cards, but it's really not that important unless they send you the additional monitors for free

The other thing I'm also interested in is how older high-end/mid-range cards compare to newer low-end cards.

Last thing I'd like, but didn't see suggested yet, is fan noise and temps when manually set to 100%. I notice a lot of cards (like mine) have fan profiles that don't even get close to 100% (mine seems to go 30-40). I notice up to a 20c drop in temps when I manually set the fan to 100, but it's also a ton louder.

I want 2048 x 1152.

I just compare to 1920x1200, it's almost the same pixel count.

If I had to guess I'd say 2560x1440 will be the next popular resolution. http://accessories.us.dell.com/sna/productdetail.aspx?c=us&l=en&s=dhs&cs=19&sku=224-8284&~ck=baynoteSearch&baynote_bnrank=14&baynote_irrank=0

Just compare to 2560x1600.

 

[HalfAHertz]

I just compare to 1920x1200, it's almost the same pixel count.

Just compare to 2560x1600.

That's true most of the time but sometimes game optimizations are coded in such a way that they work only on certain resolutions and every time you switch to an unsupported resolution, the optimizations break and performance drops. So even if they have similar pixel count, they're not the same thing.

 

[W1zzard]

That's true most of the time but sometimes game optimizations are coded in such a way that they work only on certain resolutions and every time you switch to an unsupported resolution, the optimizations break and performance drops

what is that based on? all modern games are coded to work at any resolution. pixel and vertex shaders dont know the screen resolution anyway

 

[HalfAHertz]

what is that based on? all modern games are coded to work at any resolution. pixel and vertex shaders dont know the screen resolution anyway

Mainly on limited personal observations of older titles

 

[theJesus]

Mainly on limited personal observations of older titles

Post some benches to verify your claims.

 

[HalfAHertz]

Well the only example I can think of right now is Freespace 2 MVP, which is quite an old title (the open source version of freespace 2) where I consistently get lower fps if i set it below the native resolution. Tho come to think of it that may be a CPU limitation as well.

The other game I've experienced this is Theif 2 but only on certain maps. I was playing with the resolution and FoV angles and noticed it there as well.

I'll try to do some fraps runs tonight if i have the time.

 

[W1zzard]

I was playing with the resolution and FoV angles

changing FOV changes the amount of geometry rendered

 

[qubit]

Well the only example I can think of right now is Freespace 2 MVP, which is quite an old title (the open source version of freespace 2) where I consistently get lower fps if i set it below the native resolution. Tho come to think of it that may be a CPU limitation as well.

The other game I've experienced this is Theif 2 but only on certain maps. I was playing with the resolution and FoV angles and noticed it there as well.

I'll try to do some fraps runs tonight if i have the time.

Lowering your resolution will always increase your framerate, as there's less rendering to do. The only exceptions are due to software bugs, like I've seen in graphics card reviews in rare instances.

W1zzard is right about the FOV setting. Setting a wider FOV will give it more to render and therefore slow it down.

 

[WhiteLotus]

Interactive score tables. Like you choose three or four cards and it only shows the results for those cards.

No idea how you would implement that though.

 

[Wyverex]

Interactive score tables. Like you choose three or four cards and it only shows the results for those cards.

No idea how you would implement that though.

You mean sort of like how Tom's Hardware has it?
It's quite nice, but I somehow prefer W1zzard's way.

 

[LAN_deRf_HA]

Well the only example I can think of right now is Freespace 2 MVP, which is quite an old title (the open source version of freespace 2) where I consistently get lower fps if i set it below the native resolution. Tho come to think of it that may be a CPU limitation as well.

The other game I've experienced this is Theif 2 but only on certain maps. I was playing with the resolution and FoV angles and noticed it there as well.

I'll try to do some fraps runs tonight if i have the time.

Now that you mention it that could be what people have observed with WOW. For whatever reason even if you run the game windowed (as most do) you should still set the resolution to your desktop res or else you'll take a performance hit. That never made sense to me but I noted it on quite a few builds.

As for FOV, I know weird stuff has happened with console port and aspect ratios cropping things out so that might account for some oddities in performance.

 

[W1zzard]

Interactive score tables

my main issue with those is that they cant be linked to on forums or may not work on all system (flash/js disabled)

 

[LAN_deRf_HA]

Could you put a link near the table like "get permalink" and have a picture generated for that user's config? Or would storing all those images eat up a too much space?

 

[theJesus]

Interactive score tables. Like you choose three or four cards and it only shows the results for those cards.

No idea how you would implement that though.

You mean sort of like how Tom's Hardware has it?
It's quite nice, but I somehow prefer W1zzard's way.

Yeah, those performance charts can be quite useful sometimes, like for when you can't find two cards you're trying to comparing in the same review anywhere.

my main issue with those is that they cant be linked to on forums or may not work on all system (flash/js disabled)

I think interactive performance charts would work best the way Tom's Hardware does it: Have them separate from the actual reviews and only update them like once or twice a year.