My friend Greg NI8R turned me onto a sight that I find very interesting. Unfortunately it looks like I got there a couple weeks late. It's the site of Mike N8MSA and he has some interesting ideas about ways to standardize comparison of SDR's. My opinion is the ARRL's format of testing and even Sherwood's list are not granular enough and measure the wrong things. SDR can be done in several architectures but the bottom line is what way the information is presented to the operator.
Here is a response I posted on the Flex community list regarding buying a 6300 v 6500. I own both:
"I have both radios and find the 6500 far more versatile. The 6xxx series radio is an interface between the operator and the band. As such it presents both visual and aural information to the operator. I look at the 6xxx panadapter as 2 receivers, a visual receiver overlain with a aural slice, or possibly 2 aural slices. It turns out however it presents much more data. The waterfall for example presents history and persistence. CW skimmer presents a kind of intelligence. So what the radio can present in the 6500 is far more data in more versatile ways. You can have one pan with a fast waterfall and one with a slow waterfall, you get 2 different dimensions of history. CW skimmer presents a kind of intelligence to the data it collects. 2 CW skimmers on 2 pans with 2 DAX channels do not present identical data One skimmer will decode different stations the other skimmer does not see so more data is presented by 2 skimmers So what the radio can present in the 6500 is far more data in more versatile but yet readily understandable to the op
I'm a low band DXer. Often lets say in the case of an Indian ocean DXpedition I will set up 2 pans with 2 slices in one and and once slice in the other. I will set up one slice on the DX or at least where the spot of the DX says they are, and lock that slice. I will then start using the other 2 slices to click around and start copying different waves of EU stations working the DX as they gray line approaches them. By listening to the activity I get some good idea of propagation and the skill of the operator and some idea of his pattern. As the gray line approaches me I can start to hear farthest east US ops begin to make contact. About an hour later is when they gray line finally hits me and I'm very informed about this operation. You can do this with a 6300 but it's more robust with a 6500.
I use the same setup to work very wide DX pileups like the 40khz wide variety. I set up one pan zoomed into about a 20khz wide pan with 2 slices, and a second pan zoomed out to the full width of the pileup with one slice. I will have 2 skimmers running collecting data with one skimmer tuned high in the pileup and one tuned low. I will have each skimmer set up so if I click tune on the skimmer TX falls to that slice. This way I can monitor much more skimmer data and where stations are located in the pileup. If I hear the DX calling somebody I invariably know where the station being called is in the pile up because I've been studying the pile up and I can rapidly click to that part of the 40khz pileup so my transmitter is very close to where his receiver is tuned. It also makes it easier to discern the ops listening pattern, like he makes a contact and then tunes up 500hz so tail ending is not an effective strategy, tuning up 500 or 1000 hz is an effective strategy. You get some of this with a 6300 you get more with a 6500.
I've set up a kind of SO3R for bandfils in DX contests. I'm generally not part of a contest to make the most contacts, I generally join in to give contacts and get band-fils. With writelog SDR-Bridge and skimmer I have setup my radio with 3 pans on 3 bands, 3 writelog band maps, populated by 3 skimmers and a a spot client. This mines therefore a huge amount of data on 3 bands. by linking to my spot collector database and DXkeeper which is my logbook I can easily check my needs against the 3 band maps. Quite powerful. You can do 2 with the 6300. You can do 4 but for me it becomes unwieldy
You can monitor for 6M band openings on one pan while working DX splits on another band.
I ran 2 top ten DX peditions, on 2 bands, in 2 pans, with 2 slices in each pan, driving 2 DAX channels to 2 CW skimmers, one for each DX pedition. The pans and skimmers were each setup for independent click tuning of the TX slice either in the pan or in the appropriate skimmer, and TX would follow where I clicked. My station is set up so the antennas and amp band follow the TX slice. I used a broad band RX loop for RX and switched to full duplex. So I could hear 2 top 10 DXpeditions on 2 bands in full duplex simultaneously and click tune my transmitter/amp/antenna anywhere in either of the 2 pileups. A 6500 exclusive
I use the preamp/attenuator in the 6500 extensively. I work many stations only 1-2 dB out of my noise and being able to move the dynamic range up and down rapidly in the noise to get best SNR is extremely useful. I often need to kick in some preamp when I switch to the RX loop. I have all of those preset in macros with DDUTIL so I can hit one macro and the RX loop kicks in and the preamp goes to +10 or I can hit it again and it toggles back to antenna 1 with -10dBm attenuation. I have about 5 permutations of this for various bands and antenna configurations. Some of this can be done in the 6300 but the subset of choices is inferior compared to the 6500."
The OP responded he really wanted to know would the different clock speed make a difference in how CW works in the radio. My experience is it makes no difference but some people are incredibly particular about their "CW experience" so the answer is UNKNOWN because it s the information presented to the operator that determines what's important. All of the magic I described meant nothing. So I am glad Mike took it upon himself try try and come up with a scheme to rationally compare SDR's