Discussion in 'Reviews' started by W1zzard, Mar 15, 2010.
To read this review go to: http://www.techpowerup.com/reviews/HIS/HD_5450_1_GB/
$70 for a card that can't play any modern games, or most older games...no thanks.
But it has DX11, that is a good...
Kind of makes me wonder why anyone would buy this card over one of the cheaper HD 5450's in the $40-50 range that are also low profile cards. Granted, those would only have 512MB of RAM, but what does that matter on a card of this class?
I realize that these cards are more aimed at HTPC than gaming. However, personally, if I was buying a card for a HTPC, I'd get a $20 8400GS. It will do everything I need from a HTPC card, no point in spending any more money on something better. Power consumption might be higher with the 8400GS, I don't really know how much power an 8400GS really uses, but it wouldn't be a whole lot more.
Or maybe even better, ATi can release an HD5350 with 40 Stream Processors, that is in the $20 range and consumes even less power. That would be a killer HTPC card!
This seems like a stupid card, I'd much rather pick up a 9600GT for gaming, or a HD4350/8400GS for an HTPC. Great review, nonetheless
Wouldn't be so stupid if the card was 40 bucks, because its worth that.
the card is more expensive because it has 1 GB .. our review proves that there is no point in having 1 GB, better go for the cheaper 512 MB variant
Exactly, a great review W1z, and you've kind of been proving that for years now.
I wish graphics card manufacturers would stop cramming uselessly large amounts of memory on their cards just so they can justify jacking up the price and suckering in un-educated consumers.
lol so what your saying is....you want company's to stop taking advantage of those that don't know any better?
OT: Great review as usual, useless card
I guess it's a card targeted for those who want just something better than onboard graphics.
Also reminds me of my old FX5500 with 256MB RAM. As if its gpu could ever catch up and utilise that ram...
It's still better than the 9400GT, but I'm ready to bet W1zzard facepalmed while running Crysis on it, so there was no need to try Metro 2033!
13W at full load is just, speechless. And if it is capable of running blu ray smoothly then it's fine for an HTPC.
It's all marketing. More is better. If you stick 2GB of memory on a rotten GPU and ppl will buy them like crazy. My personal experience from retail store where casual ppl buy computers, first thing they always ask is how much memory is there on a graphic card. Or they rush in asking if we have a computer with X GB of memory on a graphic card. Then, after i explain to them that the type of core is the most important thing, they finally get it. Mostly. At least good thing now is that all cards have 1GB And they are all GeForce 9500GT. They only ask about memory s it doesn't really matter Cruel reality.
believe it or not, i am 100% sure that his will sell more of this card than of the 512 mb one. everybody in the graphics card industry confirms that
Oh, I'm sure of it. I've had people, even against my advice, buy cards simply because of having more memory, even going so far as getting an HD4350 w/ 1GB DDR2 for $60 over an HD4650 w/ 512 GDDR3 for the same price...:shadedshu
I would have preferred to get the 512MB DDR3 HD4670 than the 1GB DDR2 HD4650, but it was sold out and I needed a card
What a waste of silicon. HTPC and grandma boxes only.
Mar 14, 2010, 07:00 PM
Is there something with your post date
the article has been work in progress for ages because i had to retest my cards on 10.3
aww for all the crybabys
for your work
One question W1zz - any idea why ATI would "rebadge" the memory chips on this card? Is there a possibility that GlobalFoundries is producing memory chips now?
Nice review, and good point about the memory. Honestly why does a bottom range card need as much as my 5870. The only reason it needs that is because it can handle resolutions in games higher than 1024x768, which is something this card can't do. I see the only reason to have this card is for older gaming (which can be done on 512MB or less) or for an HTPC which I would probably do when I finally get around to building one.
One thing though, why are you mentioning as a con 'No support for CUDA/PhysX'? It's like mentioning on a NVIDIA card 'No support for CCC' or 'No support for ATI Stream'. CUDA isn't something I miss since moving to ATI and I can see once they finally start making Bullet and the like work on OpenCL, CUDA will be old news...
Precisely. The thing is though, modern integrated graphics is pretty decent. I believe NVIDIA are now doing 9400GTs onboard and this isn't much better, it just does DX11 (which can't really be utilised because it's so poop). I'd only use it over integrated if the integrated was something really old, like really really old as in Intel 945.
13W at full load is impressive indeed, but its overall performance is just too low.
I'd agree on the comment - it seems weird to see a negative about hardware PhysX/Cuda, when hardware physx is so underused, and ATI has stream (even if its less widely used) - seriously, PhysX (15 game titles use it in hardware) and coreAVC aside, who even uses cuda?
my only other question, and its mere curiosity: is 3Dmark 03 still relevant for power consumption? wouldnt a more strenuous test (DX10? DX11) use more of the GPU, therefore more power - kinda like how furmark throws my wattage another 100W up compared to any game.
I agree FurMark is more stressful and that 3DMark03 probably won't be accurate anymore, but the thing is with FurMark is it's supposed to find the absolute maximum under worst circumstances. Yes it's good to know that but more often than not you will be way under it. I reckon more using 3DMark06 or Vantage to find it would be smart.
well i know why w1zz is still using it - so that he doesnt have to go rebench ALL the hardware to make the results comparable anymore.
he also cant really use a test that not all cards can use (example: DX11 only test cant be used, since not all cards could run it)
i only ask because 03 is so old, perhaps it is time to update - IF the results are massively different.
Yeah if there is no difference then probably don't bother. It should still be accurate but I wouldn't know how much.
and thus we ask w1zz to test it!
cause we know he loves retesting everything due to 2% differences in the results.
Well it might a good idea for an article: do old games use less power than new games?
Separate names with a comma.