Pages:
Author

Topic: AMD HD 6790 quickest return on investment? - page 2. (Read 3593 times)

newbie
Activity: 29
Merit: 0
Quote
One 5830 will actually pay itself off in 2 weeks.

The 5830 averages 286 Mh/s in a survey of 20 of the entries on the Hardware Mining Comparison Chart.
The 5850 averages 236 Mh/s in a survey of about the same number of entries (can't remember now).  That is an incredible difference!

I guess it can be attributed to the fact that the GPU used in the 5830 is the same one used in the 5870.

Right now Amazon has a 5830 for $154 and a 5850 for $161.  Very small diff in price, but big diff in Mh/s.
newbie
Activity: 29
Merit: 0
Going by the information at http://bitminer.info/  I noticed that the AMD HD 6790 appears to be the card whose investment one could recoup the fastest.  This is 72.91 break-even days - just over 2 months.

But in Maximum PC's July issue, in the article "Attack of the $150 GPUs," (I scanned it in and uploaded it here) the 6790 was billed as playing second fiddle to the ever-so-slightly-higher-priced 6850 with this conclusion: "In the end, [the 6970] offers decent enough performance in its class, but bear in mind, for just a few bucks more, you can pick up an HD 6850."

It says that the 6790 is derived from the Barts GPU used in the 6850, but with a big chunk disabled (READ: disabled at the factory).  The 6790 and 6850 both have the same number of transistors (1.7B) but the 6850 has 960 stream processors while the 6970 only has 800.  ROPS are cut in half on the 6970 from the 6850's 32.

To compensate for this, they bumped up the core voltage and speed, making a second power cord to the 6970 necessary (the 6850 only uses 1).  Now at https://en.bitcoin.it/wiki/Mining_hardware_comparison the Mh/s for the 6790 is listed as 220 while that of the 6850 averages 236.6 (from a low of 171 to a overclocked high of 301.4).  Since the single entry on that chart for the 6970 states the card was overclocked by 140Mhz above the factory default of 860 Mhz, it is safe to assume one would get lesser results from a typical retail card, let's say 195 Mh/s. 

I don't know what "simple maths" kristopher did over at http://bitminer.info/ but I'm sure they're accurate as long as his data is accurate.  Taking the only entry from the Mining Hardware Comparison Chart and using it to base the rest of his data on is fine - it's the only data there - but in this case it is misleading.

I don't think the 6970 is a very good card to buy for mining based on what I've seen thus far.  If anyone around here is using these cards, please update the Chart so we can have more data on them.
Pages:
Jump to: