Discussion Thoughts on the Radeon HD 5850?

[jnjnilson6]

I have never had a 5850. Although I've had a friend who'd owned it. He had been financially tricked into getting an AMD 1090T or 1100T (forgot which), 4 GB RAM and an HD 5850 in 2010 for over 3000 dollars. Well, the point of this writing is not that. Many people are continually saying the 5850 was only for very old and simple games, yet it was faster than the HD 6850 and the HD 7770 and the GTX 560 SE . It could play Crysis and games like BF4 at 1920x1080. The graphics settings would be pretty high and the framerate - pretty good.

Well, why do you think that is? Why do people think of an enthusiast and revolutionary GPU (for its time) as something very weak worthy of conquering only light titles? Back in its day 1080p was the thing and it did damn good.

I was just wondering why the true power of this beast of a card (for its time) is kept hidden and many people have got a wrong impression of its performance (saying it was meant for something like Far Cry 2); the 5850 GPU marking a line of distinction between the cards which vaguely tried to be 'breakthrough' and the serious cards which brought entirely new performance to the sphere of gaming. Dazzling sunsets and cars scurrying over forgotten Mesas, the framerate punching through innumerable pixels and a heavy, sturdier presence being spotted within.

What do you think?

 

[Darkbreeze]

It's not about the resolution, it's about the demand of the games. The games that existed when that card was new and could be played at 1080p were no where near as demanding as games that were current and popular at the time those later cards came along, or now.

For the record, the 6850, 7770 and GTX 560 SE were only marginally better than the 5850 anyhow. So the reality is that ALL of those games are only good for very old and simple games, compared to anything that was actually capable for that time like the HD 6990, 7990 or GTX 580/590.

Even for it's time, the 5850 was never a "beast" in any sense of the word.

 

[jnjnilson6]

It's not about the resolution, it's about the demand of the games. The games that existed when that card was new and could be played at 1080p were no where near as demanding as games that were current and popular at the time those later cards came along, or now.

For the record, the 6850, 7770 and GTX 560 SE were only marginally better than the 5850 anyhow. So the reality is that ALL of those games are only good for very old and simple games, compared to anything that was actually capable for that time like the HD 6990, 7990 or GTX 580/590.

Even for it's time, the 5850 was never a "beast" in any sense of the word.

You're right, of course! I wrote about those cards because my friends had them at the time and through a variety of different games they portrayed synonymous performance, although I think that the 5850 somehow had a fairly relative and little, but still countable performance gain. The guy with the 6850 had an i5-2400, the one with the HD 7770 - I think he had a good i5 3rd gen and the one with the 560 SE - a Core 2 Quad (I think Q8200). At the time I had an i7-3770K and HD 6770 (Sapphire). So I was just a little bit behind them in most games; their CPUs were really good for the time, counting out the Core 2 Quad.

Again, I fully support your opinion, but still there's that little bit of wistfulness for those heightened numerals - 5850, a subtle margin before the 5870.

 

[Darkbreeze]

I think that the 5850 somehow had a fairly relative and little, but still countable performance gain.

Over those other adapters? No, not really. Each of those that you mentioned got incrementally better.

 

[jnjnilson6]

Over those other adapters? No, not really. Each of those that you mentioned got incrementally better.

Gathering up the times from memory, as I'd been an indefinitely avid 'hardware geek' at the time, I do seem to remember that the 5850 actually did beat the 6850, the same way the 5770 beats the 6770.

 

[jnjnilson6]

Over those other adapters? No, not really. Each of those that you mentioned got incrementally better.

And this is from https://www.techpowerup.com/gpu-specs/radeon-hd-5850.c252 (TechPowerUp).

 

[jnjnilson6]

Over those other adapters? No, not really. Each of those that you mentioned got incrementally better.

Here again if we count the GTX 560 SE (TechPowerUp).

Well, I remember back in the day thinking (throughout the HD 6xxx, 7xxx and GTX 5xx and 6xx times) that CPUBenchmark really held it up and well in benchmarking, until on the GPU section I came across multiple exceptionally wrong entries concerning the cards on top and other cards scrolling along; this really shook me because I really had trust in the site and afterward I stopped using it as a pitstop of performance recollection. I had found on the CPU side of the site discrepancies too, but not so large and blinding as the aforementioned.

 

[Darkbreeze]

Userbenchmark is useless. Literally a running joke amongst all veteran builders and enthusiasts that I know, and don't know.

"Relative" performance is percentage ranked performance based on cost. It's performance RELATIVE to cost, not pure performance, so no, the Passmark scores are not wrong.

 

[jnjnilson6]

Userbenchmark is useless. Literally a running joke amongst all veteran builders and enthusiasts that I know, and don't know.

"Relative" performance is percentage ranked performance based on cost. It's performance RELATIVE to cost, not pure performance, so no, the Passmark scores are not wrong.

That's what I'd thought back in the day; but there were some mind-striking discrepancies in some of the listings which seemed so profound as to be either hilarious or doubtfully boggling the contemplation in the somber hours as to why and how they managed to get them there. They were literally blinding, appearing boldened and indefinite within their inconsistency, like a bad joke in the morning. It was literally something to ponder about at three o'clock after midnight - the stability and resourcefulness of the site ended blearily for me past the point of those mistaken listings. It could be better today, but that was the turning point. Incorrect and dragging away the contemplation at a slant, what I'd come to note made my virtuous and harmless enjoyment in benchmarks turn a whitened corner into other sites and other benchmarks. There are other people whom this is familiar to, people who have changed their mind in regard to the plausibility and correctness of the site.

Thank you for writing up and sharing your opinion; it was a truly invaluable asset in discussion!

 

[Darkbreeze]

😎

 

[jnjnilson6]

😎

This is what I meant. I used the Wayback Machine to go back in time and as you can see the site shows the 6870 to be slower than the 6850 which should be fully impossible.

 

[jnjnilson6]

😎

It also shows the 6950 to be faster than the 6990.

 

[jnjnilson6]

😎

It even shows both the 6990 and 6970 as slower than the 6950 (as per the screenshot from the previous comment).

This is pretty much the reason why I'd stopped trusting the site I had trusted most at one particular point in time. And I haven't really come back because in those years there were some results so ridiculous even a 10 year old kid would find them fault.

Now, it is good what you have provided, and I am almost certain the site is better run nowadays. But everybody's got to agree that once there were surely incorrect results and we may be only so certain that there aren't some veiled deviations nowadays too underneath the fancy exterior. People take the site for granted as I used to in the deep past, but there is obviously a frayed cuff here and there and a mischievous gleam beneath the bright white polish.

 

[Darkbreeze]

Passmark in 2011 ain't Passmark in 2022. Everything "back in the day" was suspect. It's ALL much more refined now. Perfect? No, of course not. But the mere fact that Passmark HAS been around for a long time and has some of the more respected products on the market for bench testing for at least the last five years or so says something. And, you also have to factor in that they can't account for ALL of the variables so there are going to be some discrepancies sometimes.

Every test result isn't going to be exactly the same, no matter WHAT site or utility you want to talk about because there will always be differences in CPU scores when the systems being tested have different graphics cards, different memory capacities and speeds and latencies, different motherboards (And yes, the model of motherboard involved in ANY kind of testing makes a difference. Sometimes a BIG difference), different drive speeds and so on, and equally when testing the GPU performance, those factors have some weight as well.

Same graphics adapter but two different CPUs, are going to have different results. The only way to ever have a completely level playing field is to have ALL identical hardware, and even then, there will still be some variance in the quality and binning of the hardware.

 

[jnjnilson6]

Passmark in 2011 ain't Passmark in 2022. Everything "back in the day" was suspect. It's ALL much more refined now. Perfect? No, of course not. But the mere fact that Passmark HAS been around for a long time and has some of the more respected products on the market for bench testing for at least the last five years or so says something. And, you also have to factor in that they can't account for ALL of the variables so there are going to be some discrepancies sometimes.

Every test result isn't going to be exactly the same, no matter WHAT site or utility you want to talk about because there will always be differences in CPU scores when the systems being tested have different graphics cards, different memory capacities and speeds and latencies, different motherboards (And yes, the model of motherboard involved in ANY kind of testing makes a difference. Sometimes a BIG difference), different drive speeds and so on, and equally when testing the GPU performance, those factors have some weight as well.

Same graphics adapter but two different CPUs, are going to have different results. The only way to ever have a completely level playing field is to have ALL identical hardware, and even then, there will still be some variance in the quality and binning of the hardware.

Yeah, you're perfectly right about that! I've experienced it too back in the day while benchmarking; ulterior particulars of the configuration may well make the score of the tested part deviate by up to 15%. So, as you've been careful to explain, we may not have an exactly synonymous environment everywhere and there is always space for deviation here and there.
This is what I've been thinking too, of course. But it was very well said and very properly explained!

It does show of course that you've been working a lot with hardware. I used to be atop the crest of the wave before some years now, however, it has turned into a senile passion, shaded vaguely and reminiscently in the past, replaced by a different medium. The discussion turned out positively satisfactory, as it obviously would have when talking with a professional.

 

[jnjnilson6]

Passmark in 2011 ain't Passmark in 2022. Everything "back in the day" was suspect. It's ALL much more refined now. Perfect? No, of course not. But the mere fact that Passmark HAS been around for a long time and has some of the more respected products on the market for bench testing for at least the last five years or so says something. And, you also have to factor in that they can't account for ALL of the variables so there are going to be some discrepancies sometimes.

Every test result isn't going to be exactly the same, no matter WHAT site or utility you want to talk about because there will always be differences in CPU scores when the systems being tested have different graphics cards, different memory capacities and speeds and latencies, different motherboards (And yes, the model of motherboard involved in ANY kind of testing makes a difference. Sometimes a BIG difference), different drive speeds and so on, and equally when testing the GPU performance, those factors have some weight as well.

Same graphics adapter but two different CPUs, are going to have different results. The only way to ever have a completely level playing field is to have ALL identical hardware, and even then, there will still be some variance in the quality and binning of the hardware.

Something interesting I did find. In the instances herewith the 5850 seems to take over both the 6850 and 7770 in terms of gaming performance. It seems to beat on at a higher framerate, computing power portrayed herein.