Sign in with
Sign up | Sign in
Your question

Aren't you happy the midrange cards don't suck anymore?

Last response: in Graphics & Displays
Share
October 31, 2008 8:07:14 PM

I hear a lot of people call these "budget cards" making them sound very bad but I perfer to call them "midrange" which is what they really are, anyway it's been awhile since the cards with a "6" in them as the second digit have been any good.

Finally Both Nivida and ATI have cards that don't totally suck in the sub $200 price range, such cards that come to mind:

9600GT, 9600GSO, some 8800GTs and some 9800GTs.

4650,4670,4850, and some old 3850s.

I'm so glad they finally realised that is where most of the money is and decided to make them finally not suck like the 2600 and 8600 series did.
October 31, 2008 8:12:18 PM

The 8800gts g92's and 9800gtx/gtx+'s are also in the <$200 range and they're pretty decent.
a b U Graphics card
October 31, 2008 8:23:10 PM

br3nd064 said:
The 8800gts g92's and 9800gtx/gtx+'s are also in the <$200 range and they're pretty decent.


I totally agree. My 8800GTS (G92) never gets talked about anymore. Harder to find them for sale online anymore either. Great card too! Plays FarCry 2 @ Max settings @ 1280x1024 at 25-45FPS depending on the environment I'm in. :) 

Wish I could afford to go SLI with mine. ;-)
October 31, 2008 8:30:43 PM

The 4670 is impossible to find at a fair price in Canada, so as far as I'm concerned it doesn't exist or change anything. Yes, you've got affordable 8800s etc, but you need to invest in a decent power supply to capitalize on that, so overall you're just treading water. The midrange is better because everything is better, but I wouldn't say that it's relatively cheaper when you include availability and other costs into the equation.
October 31, 2008 8:59:16 PM

Does midrange normally eat more then 100w ? Its scary where these power consumption go but again the power performance ratio is mad compare to what we had so far on the market.
October 31, 2008 9:14:17 PM

IMHO, we've had it pretty good since the 6600GT or so. Don't get me wrong, it's gotten better and better except for that hickup at 8600, but, we've had it pretty good for a while now. Same with CPU's. You can get more than most people need in a $50 CPU these days. Heck, there are few games that a $50-100 CPU and $100-150 GPU can't handle at at least reasonable settings now.
October 31, 2008 9:37:42 PM

One thing that's really helped with midrange cards is that fact that games haven't been getting more graphically draining. After crysis came out things seemed to have leveled out and even back pedaled. It may be 2 years before we see another crysis level performance hog, the development costs and lack of appeal just won't allow it... even now highend cards are really only needed for filters and excessively high resolutions. Ati and Nvidia are just going to have to adjust to the fact that the market is all about price performance ratio now. No more of this 8600 GT for $200 dollars crap. People are going to get use to playing games like UT3 at full res/graphics settings for under $100 and that's just how it's going to be from now on.
a b U Graphics card
October 31, 2008 10:07:20 PM
a b U Graphics card
October 31, 2008 10:09:11 PM

And prices have been awesome over the last year.

Sure it was more about deals on the high end than quality in the mid-range, but the GF9600GT and now HD4670 have changed that.
October 31, 2008 10:17:52 PM

yeah the 8600's killed it....



like they got hard and then someone chopped with a knife
October 31, 2008 10:20:25 PM

The 8800GT really opened the door for REAL midrange in recent memory. I guess you could say the GTS G92 followed suit.
October 31, 2008 10:27:15 PM

TheGreatGrapeApe said:
OMG, it's so tough to find :heink:  ;

Seems alot of them out there, no?
So under $100 isn't good enough?
Heck, new midrange cards used to be $200-250+ !


Those are garbage prices for the 4670. Might as well just get an 8800 or 9600. It makes sense at around $75.

And newegg.ca? You are joking right? You do realize they ship UPS cross boarder for $17 and you end up paying customs on top... It's a US operation with a .ca domain. Screw them.
October 31, 2008 10:31:27 PM

metrazol said:
Those are garbage prices for the 4670. Might as well just get an 8800 or 9600. It makes sense at around $75.

And newegg.ca? You are joking right? You do realize they ship UPS cross boarder for $17 and you end up paying customs on top... It's a US operation with a .ca domain. Screw them.


Lol, yep, key word is "fair price." :p 
October 31, 2008 10:56:24 PM

This is exactly why competition between ATI and Nvidia is a great thing.
a b U Graphics card
October 31, 2008 11:09:08 PM

metrazol said:
Those are garbage prices for the 4670. Might as well just get an 8800 or 9600. It makes sense at around $75.


Nah, the GF8800 and GF9600 are old products sellign for low prices and you're comparing to newer cards selling near launch prices.

Might as well get a free GF2MX, it's a better value! :pt1cable: 

Quote:
And newegg.ca? You are joking right?


You said it's impossible to find them, and that's not true. I don't care even if they are re-shipped 3 other options are from Canadian distributors if you have an issue with the US.

Your idea of a fair price is also bent.

It may not be the best value, but it's definitely a fair price. Especially when you compare them to their legacy price for what is typically mid-range. The GF9600 wasn't at that price and neither were the GF8600GTS and HD2600XT which would be the equivalent of these.

Whine all you want, but they are out there at a fair price, you're statement is a stawman that you're never going to agree about a 'fair price' because if they were selling for $75 you would say they should be $50, etc. :pfff: 
October 31, 2008 11:17:28 PM

8800 GT did a good job showing as what midrange means (9600 GT came out a few later, but by the time it came out, it was like 30$ cheaper than the 8800 GT)

I think the 4850 is the new card that stepped it up offering 10-30% increase over the 8800 GTX for under 200$ on release.

We also have the GX2, 260 and 4870 512 closing the gaps now being sold for 200-250$:)  which is amazing.
October 31, 2008 11:18:26 PM

my last midrange card (before my HD4850 which is absolutely amazing) was an ATI Radeon 9600Pro, which was about $100 and it was able to play Doom 3 at great frame rates...at that time, Doom 3 was the current day Crysis...
October 31, 2008 11:24:23 PM

my friend bought the 9600 Pro socould play serious sam with me:D  I had a 9800 XT :p  lol

In all honesty my favorite "mid range" (actually semi- High end) was the X1950 XTX that I bought from future Shop (Canadian Best buy) for 150$ because I photoshoped a flyer and asked them to match the price:D  LMAO!!!!
a b U Graphics card
October 31, 2008 11:46:12 PM

However, remember the R9600Pro launched at $149-200 back when the deals weren't quite so grand. In Canada that was closer to $175-$250 (I got one from the ATi online store for just over $ 225 first week). The R9500Pro that came before it was $179-$219US and a cut down high end card, similar to the GF8800GT, not the GF9600 or HD4670 or HD2600/GF8600.

Prices have done well recently, and they trended downward from the days when the mid-range GF3s, 8500LE and GF4s were closer to $300. It'd be interesting to see how they would've done if the HD48xx series had been priced a little higher, in line with it's vlaue to the GTX2xx rather than setting a low point and dragging the GTXs down to there. Buying an HD4850 for what people were paying for an R9500/9600/FX5600 is a pretty nice price drope for class of card/power, and paying the price for a GF9600 or HD4650/70 that you would've paid for an FX5200/R9000, GF6300/X300 or X1300/GF7300 is pretty nice too. :sol: 
October 31, 2008 11:53:03 PM

The artificial tag "new tech" or "old tech" means little. Bottom line is, if old tech outperforms new tech for cheaper, it's better value. Neither do prices in the past matter. As a consumer, why do you care for past prices? It's not like you're buying in the past. :p 
http://www.techpowerup.com/reviews/Powercolor/HD_4670/2...
a b U Graphics card
October 31, 2008 11:59:53 PM

And whats this requiring a 100 watts? The 4670 doesnt need that much at all, and its a great low end-midrange card
November 1, 2008 12:00:54 AM

yup, its out performed by the 9600 GT alot, but its on avg 30-50$ cheaper.
a b U Graphics card
November 1, 2008 12:06:29 AM

dagger said:
The artificial tag "new tech" or "old tech" means little. Bottom line is, if old tech outperforms new tech for cheaper, it's better value.


Has nothing to do with a 'fair price' though. And value is all in the eye of the beer holder. You value one thing, someone else values another.
The HD4670 is a much better 'value' because it's a better HTPC card, and that's all that matters... or do you value something else differently?

You can do your whole 'value' thread again, but try and stick to the OPs idea again that it's nice that the mid-range doesn't suck, not that you have value metrics for the nth time. :sarcastic: 

Quote:
Neither do prices in the past matter. As a consumer, why do you care for past prices? It's not like you're buying in the past.


Read the point of the thread, of course the past matters and of course it's relevant to the launch of mid-range products, of which the GF8800 NEVER was. It's not just a value thread, it's a mid-range cards not sucking for either ATi or nV' thread. That's what the OP is on about if you bothered to focus. :kaola: 
a b U Graphics card
November 1, 2008 12:07:15 AM

Past prices are what this whole threads about. If ATI hadnt come in low like Ape mentioned, wed still be paying huge prices for old tech, and huge prices for new tech. Now, that being said, which would you rather have, old or new? DX10 or DX10.1? 7 channel or not? Lower power consumption or not? These are questions people want answers to, and as far as this thread, the pricing is everything, and again, thanks only to ATI.

I see your point dagger, but youre beating a dead and since renamed 3 times horse
November 1, 2008 12:07:21 AM

JAYDEEJOHN said:
And whats this requiring a 100 watts? The 4670 doesnt need that much at all, and its a great low end-midrange card


4670 doesn't require nearly 100 watts, even at full load. Since it's technically impossible to directly measure gpu power consumption, benchmarkers list whole system consumption instead. It's misleading if you're not careful.
http://www.techpowerup.com/reviews/Powercolor/HD_4670/2...

L1qu1d said:
yup, its out performed by the 9600 GT alot, but its on avg 30-50$ cheaper.


It definitely does not outperform 9600gt. It performs under 9600gso, which is cheaper and overclocks far better.
http://www.techpowerup.com/reviews/Powercolor/HD_4670/2...
November 1, 2008 12:17:48 AM

TheGreatGrapeApe said:
Has nothing to do with a 'fair price' though. And value is all in the eye of the beer holder. You value one thing, someone else values another.
The HD46700 is a much better 'value' because it's a better HTPC card, and that's all that matters... or do you value something else differently?

You can do your whole 'value' thread again, but try and stick to the OPs idea again that it's nice that the mid-range doesn't suck, not that you have value metrics for the nth time. :sarcastic: 

Quote:
Neither do prices in the past matter. As a consumer, why do you care for past prices? It's not like you're buying in the past.


Read the point of the thread, of course the past matters and of course it's relevant to the launch of mid-range products, of which the GF8800 NEVER was. It's not a value thread, it's a built for mid-range cards not sucking. That's what the OP is on about if you bothered to focus. :kaola: 


For gaming, in which context this thread was made, "value" means average fps/$ ratio. For watching movies, a $30 low range card will do fine, and you won't notice the difference with a mid range card.

When a card performs like a mid range card and costs like a mid range card, it's a mid range card. You can't exclude better performing and lower costing competition until only your favorite is left. That's your problem, you've already made up your mind before considering all alternatives, and merely try to justify your existing prejudices in whatever ways you can. :p 
a b U Graphics card
November 1, 2008 12:21:51 AM

dagger said:
Since it's technically impossible to directly measure gpu power consumption...


You're joking right?
If you think that's technically impossible, I think you need to take a basic electricians course.

It's easier to test power consumption of a whole PC, but it is SOOOoo innacurate on SOOoooo many levels, heck the power draw at the plug isn't even the same as the power draw past the PSU.

Other people do a better job of isolating the power draw of the card rather than the power draw of other items which combined draw many times more than the more efficient GPUs. Xbit does a better job of trying to isolate the cards themselves by measuring the draw at the power points, not guessing at the entire system;

http://www.xbitlabs.com/articles/video/display/powercol...

Neiter appears to draw 100W, but they aren't equal either.
November 1, 2008 12:22:31 AM

read my post again, I said that the 4670 is out performed by the 9600 GT:) 
November 1, 2008 12:32:02 AM

L1qu1d said:
because I photoshoped a flyer and asked them to match the price:D  LMAO!!!!


Awesome.
November 1, 2008 12:34:26 AM

yeah it is but watch out if they aren't lazy and they check it, they can hold fraud against you:p . I was under 18 so :p 
November 1, 2008 12:35:58 AM

TheGreatGrapeApe said:
You're joking right?
If you think that's technically impossible, I think you need to take a basic electricians course.

It's easier to test power consumption of a whole PC, but it is SOOOoo innacurate on SOOoooo many levels, heck the power draw at the plug isn't even the same as the power draw past the PSU.

Other people do a better job of isolating the power draw of the card rather than the power draw of other items which combined draw many times more than the more efficient GPUs. Xbit does a better job of trying to isolate the cards themselves by measuring the draw at the power points, not guessing at the entire system;

http://www.xbitlabs.com/articles/video/display/powercol...

Neiter appears to draw 100W, but they aren't equal either.


That's not right. Note the term "directly measure." Whole system consumption is a fact, measured from a single point. All you have to do is attach a standard meter through the outlet, and get the whole system, plus psu loss and everything. The gpu consumption is an educated guess done through deduction. Multiple channels of power. It's not just a pcie connector from psu, power from pcie bus can't be cleanly isolated. Not to mention most benchmarkers just take the difference from bare bone system and take whatever gap in consumption as what the card takes, instead of using custom meters. Hardly precision measurement.

L1qu1d said:
read my post again, I said that the 4670 is out performed by the 9600 GT:) 


Lol, good catch. +1 :na: 
a b U Graphics card
November 1, 2008 12:43:19 AM

Using no sound cards, what other huge draws would be on the PCI bus? You can measure the PCI, and also, the connector, if there is one. I dont see why you couldnt
November 1, 2008 12:53:45 AM

JAYDEEJOHN said:
Using no sound cards, what other huge draws would be on the PCI bus? You can measure the PCI, and also, the connector, if there is one. I dont see why you couldnt


Of course it'll be basically on-mark, barring some tiny consumption difference from bus and chipset behavior under the different conditions. Still, it would be a deduction, and not direct measurement, as with whole system consumption. Which fits into what I said about it being "technically impossible to directly measure gpu power consumption." Ape replied to it as though that statement is wrong. :p 
a b U Graphics card
November 1, 2008 1:01:29 AM

Dagger, once again you miss the point. It's not about X or Y being a better value versus each other. The OP isn't talking about 0.5 fps/$ he's talking about them all being better than the crap we had in the GF8600/HD2600 days. That you miss that shows you're the one with the pre-conception that everything is about your value metrics where cut./paste your argument and change the #s to match whether it's the GF9600 or GX2 you're arguing about this time.

So simple question for you since you don't seem to get it. Is the HD4670 a better card for the mid range than the GF8600 and HD2600 or do you think they "suck like the 2600 and 8600 series did."

You seem to miss the point of the excercise.

Or is you argument that the prices I posted are not fair compared to these;
http://www.ncix.com/products/index.php?sku=25451&vpn=EA...
http://www.ncix.com/products/index.php?sku=26176&vpn=EN...

Personally I'd say the HD4670 is fair, while those two.... uh.... not so much.


As for your statements about power measurements, your saying the power numbers on the Xbit test (which are actually pulled from the PCie slot and connectors "- we measure the current in shunts that we connect into the +3.3V and +12V power lines of the mainboard’s graphics slot. Similar shunts are also connected into the alternate power connector (the additional power connector on the graphics card’s PCB).
- "
) thus isolating it to the graphics slot and power connectos are more of a guess than taking the entire system at idle and under load with two different cards and subtracting them?

Sure if you want to find out how much your power compan is going to charge you for your PC running either card, that might be revelant, as for the power draw of the card itself as it would relate even to the article you posted's mention of cooling, it's a pretty wild guess. Let alone it's effect on the delivery of power from a questionable PSU.

Which type of testing do you think the IHVs would use to determine TDP, the card and chip in isolation or the entire system they are testin on and then trying to work backawards?

Interesting though, that might explain a few burnt chips.


dagger said:
Which fits into what I said about it being "technically impossible to directly measure gpu power consumption." Ape replied to it as though that statement is wrong. :p 


So to get this straight, you STILL think it's technically impossible, and that for a card like the HD4670 without extra power from thePSU, it is impossible to make a riser card that would test the power draw at the pins and thus isolate only the power draw across that space? And that for other cards using in-line meters for the molex connecors etc would let you test that either, and the system consumption would still be more accurate?
November 1, 2008 1:23:06 AM

Ape has a good point here. If it was "impossible" to read/test the draw at the pins, then its safe to say the raitings the manufactures give us are complete bullcrap. I dont buy that, thats like selling a car without telling the MPG, or not even bothering to test what it is.

The power draw can be read, and pinpointed very accurately.
November 1, 2008 1:27:05 AM

If the hd4670 doesn't have any extra power connectors then it can't be using over the 75w provided by the pci-e slot.

As for $100 midrange cards, this has always been the case. I bought the following:

1. tnt pci for $60, top end was tnt2u for $250
2. geforce2mx for $75, top end was geforce2u for $599(!)
3. geforce4 ti4200 64mb for $99, top end was 9700pro for $399
4. Radeon 9800pro for $99, top end was x800xt for $499
5. Now I bought a hd2600xt for "very cheap" 2 weeks ago. This should outperform a 9800pro, if not im looking to get an x1950pro for a few $ more thatll be way faster. It remains to be seen, I am going to test that hd2600xt soon. I have windows xp so I can't use dx10 anyway.
November 1, 2008 1:41:40 AM

i had a 6600gt, a radeon 9600xt, and a7600gtoc(one of my favorite cards) and those definately did not suck, i actually was playing bioshock on the 7600gt when i went quad core, and it still looked good(not great like my 9600gt, but good). All i will ever buy is midrange cards, i will evers spend over 200 dollars or do sli or crossfire, no need imho
November 1, 2008 1:57:22 AM

reconviperone1 :

I thought the main point of the thread is not that those (Although the thread title is a bit misleading) video cards suck in general but the kind of improved performance you would expect in the midrange segment. I mean the performance difference between the 8600GT and 9600GT and the Radeon 3650 and the Radeon 4670 are really huge to say the least. And they sort of cost similar as well, if not lower (If adding inflation.)
November 1, 2008 10:42:12 AM

So according to dagger even if the card is eating 140w to 170w of power its a perfect midrange even if a 4670 thats about 20 to 30 percent slower and uses almost a third of that power consumption region the 88s gets a thumbs up. Thats great for the budget gamer who need to throw in liquid cooling for sli seeing fans wont cool those monsters don't forget a 850w powersupply to keep that htpc stable.
November 1, 2008 5:16:34 PM

My last card was a 7900GS and I got that on sale for $180. I got my 4850 for $200 and its way better for $20 more 18 months later.
a c 376 U Graphics card
November 1, 2008 6:23:42 PM

cyber_jockey said:
So according to dagger even if the card is eating 140w to 170w of power its a perfect midrange even if a 4670 thats about 20 to 30 percent slower and uses almost a third of that power consumption region the 88s gets a thumbs up. Thats great for the budget gamer who need to throw in liquid cooling for sli seeing fans wont cool those monsters don't forget a 850w powersupply to keep that htpc stable.

Overstate things much? The 8800gt has a real power consumption of around 80 watts and everything I've seen suggests the 2 9600s are within about 20 watts of an HD4670 under load.
The HD4670 is a nice card but in terms of gaming performance and price it's simply not the best card for the money at the moment unless your PSU is craptacular.
November 1, 2008 7:00:16 PM

The 98/88gtis without a doubt a good midrange graphics card but its cherry has pop with big bad 4830 around. As for the 4670 its simply attracting another crowd. Bluray anyone ?
November 1, 2008 7:50:31 PM

They are basically even, the 8800/9800GT and the 4830. All it comes down to is price and brand loyalty when picking between them.
November 1, 2008 9:12:58 PM

Nvidia cards there cracking value, more so than ATI IMO. Nvidia produce drivers of higher quality than ATI usually, Game development support is also typically larger. You know when you buy an Nvidia based card, that most games will be fine tuned to run optimally on your new hardware. The problem with ATI is they constantly play the catch up game because they just don't have the development support, which Nvidia has. What you really want to know is, what does this actually all mean to the average Gamer? well optimized games suffer less from graphic corruption and performance related issues. ATI may have fast hardware, it's just the hardware cannot perform to it's true potential because optimizations are poor and driver support lags behind Nvidias.

I believe Nvidia offer the stronger products and because Nvidia work closely with many game developers, means the experience will typically be better. I don't hate ATI BTW I just believe no one should settle for second best, with their hard earned money. Nvidia is 'The way it's meant to be played' quite literally.

What I want to know is, why hasn't ATI fixed those heat related issues with 4000 series? These Cards are running uncomfortably hot by default. The way to fix the overheating issues is to change the fans manually, it's shame the unaware buyers could potentially face premature hardware failure because of this and don't let anyone tell it different, no justification in hardware running 80-90 something degrees. I think this is an issue ATI really need to address and stop threating their customers so poorly.

I'm not a fanboy, just believe in better products.
a c 376 U Graphics card
November 1, 2008 11:24:55 PM


Did you actually look at the reviews in that article or just want to post a headline from a ATI fanboy site?
Despite the headline 3 of the 4 reviews I picked at random from the article had the 9800gt equal to or beating the HD4830 in most benchmarks.
The 4670 and 4830 are nice cards but at this point they are quite simply being undercut by nvidia in terms of price/performance. Maybe the prices will change but for the moment that's reality even if you don't like it. Flat-out lying about power consumption and linking to articles with biased headlines wont change that.
a b U Graphics card
November 2, 2008 12:17:31 AM

Ummm, the 4830 beats the 88gt, not by much, but it does, maybe 5% overall, but it still beats it. I HAVE looked at Annands, Tech Report etc, and thats their conclusions
November 2, 2008 12:47:07 AM

Well jy jy I simply took the first link with multiple benchmarks across the board and as fanboyish as that title is I think that 4830 does completely wipes the floor with the 98gt. Overall its down right faster not by much but faster more features dx 10.1 and the list goes on.
!