Welcome to Vizjerei Clan Forums

Register now to gain access to all of our features. Once registered and logged in, you will be able to contribute to this site by submitting your own content or replying to existing content. You'll be able to customize your profile, receive reputation points as a reward for submitting content, while also communicating with other members via your own private inbox, plus much more! This message will be removed once you have signed in.

The conversion to the new forums and skin is done. Information, as well as to leave feedback and comments can be done so by going to thread: [NEW FOR 4.1.X FORUMS] - Feedback & Support

dolynick

Former Officers
  • Content count

    1,581
  • Joined

  • Last visited

About dolynick

  • Rank
    The 'nadian

Profile Information

  • Gender
    Not Telling

Recent Profile Visitors

1,516 profile views
  1. Domain ownership transfer in-progress.  Not sure if there will be any site downtime while it's happening or not.

    1. void1872

      void1872

      whos owning it now?

    2. dolynick
  2. Also 0 knowledge of "R" but... If you want to sort two arrays that are not somehow linked in some sort of structure, then I'd expect that you would need a third array that consists of ordinal order for display. That third array would basically be a sorted list of value indexes from your medal count array and then you would apply that array as your plot order. Basically if you had an array of medal counts of 1, 10, 2, 5 then your third array would be 1, 3, 4, 2 and you would plot sport[1]+oly[1], sport[3]+oly[3], sport[4]+oly[4] and sport[2]+oly[2] (IE, indexes of the two data arrays according to the values in the 3rd "sort order" array). This example is in ascending order and you want descending but same idea. You won't be able to sort the two arrays independently because even if you could sort an array of strings (which I'd expect to throw an error if it was some sort of numerical sort command), its results wouldn't match the array of integers for medals in any way, shape or form. That's why you need the 3rd array. This would also assume that the indexes of both your data arrays correspond and that both are the same length (which they probably are for the exercise's purposes). Then again, not knowing anything about R, it may function at a high enough level that you aren't dealing with that level of minutia. Most of my scripting/coding time has had nothing to do with plotting outputs and sample code you posted is completely foreign to me in terms of its syntax and formation. I could write you a pseudo-code bubble sort and all that (as I'm sure Void could too) but I'm not sure that would be of any help or relevance here. If you do have some of structured array linking the two arrays (IE structure[1] contains both oly[1] and sport[1]) then you could sort them in one go or command if such was available. All you'd have to do is sort structure by its structure.oly values.
  3. As far as I know, the breakpoints and the calculations in the WizTools are still correct. Mammoth Hydra: Min APS - Max APS - Tick/s - Length/t - Frames/t - IAS to BP 10.94225024 76.59574468 10 0.1 6 671.5893029 5.89198136 10.94224924 5 0.2 12 310.8558114 4.031355983 5.89198036 3.333333333 0.3 18 177.9539988 3.063830787 4.031354983 2.5 0.4 24 108.8450562 2.470831474 3.063829787 2 0.5 30 66.4879624 2.070156262 2.470830474 1.666666667 0.6 36 37.8683044 1.781297388 2.070155262 1.428571429 0.7 42 17.23552771 1.563179463 1.781296388 1.25 0.8 48 1.65567592 1.392650903 1.563178463 1.111111111 0.9 54 -10.52493548 1.255668946 1.392649903 1 1 60 -20.30936103 1.14322107 1.255667946 0.909090909 1.1 66 -28.34135215 1.049257776 1.14322007 0.833333333 1.2 72 -35.05301597 0.969567388 1.049256776 0.769230769 1.3 78 -40.74518655 0.901127408 0.969566388 0.714285714 1.4 84 -45.63375657 0.84171248 0.901126408 0.666666667 1.5 90 -49.87768 Other Hydras: Min APS - Max APS - Tick/s - Length/t - Frames/t - IAS to BP 10.90281453 76.31969472 10 0.1 6 668.7724665 5.870746748 10.90281353 5 0.2 12 309.3390534 4.016827038 5.870745748 3.333333333 0.3 18 176.916217 3.052788789 4.016826038 2.5 0.4 24 108.0563421 2.461926636 3.052787789 2 0.5 30 65.85190258 2.062695452 2.461925636 1.666666667 0.6 36 37.33538942 1.774877621 2.062694452 1.428571429 0.7 42 16.77697296 1.55754579 1.774876621 1.25 0.8 48 1.253270731 1.387631813 1.55754479 1.111111111 0.9 54 -10.88344192 1.251143536 1.387630813 1 1 60 -20.63260454 1.139100921 1.251142536 0.909090909 1.1 66 -28.63564848 1.04547627 1.139099921 0.833333333 1.2 72 -35.32312356 0.966073085 1.04547527 0.769230769 1.3 78 -40.99477964 0.897879761 0.966072085 0.714285714 1.4 84 -45.86573133 0.838678964 0.897878761 0.666666667 1.5 90 -50.09435972 Sorry for the formatting, it doesn't come over from the tools all that well unless I go to the trouble of exporting as space delimited and all that crap. Ignore the last column of IAS to BP since that is being calculated form the default sheet APS that was in the WizTools when I copied it (it was 1.54 APS FYI). The reason for the difference is because Hydras have an Base FPA (frames per action) of 2.3585/3 whereas Mammoth has a Base FPA of 2.35/3. If you want to calculate a breakpoint for a hydra use these formulas Bottom of BP range = 60/FPA/(frames+1)+0.000001 Top of BP range = =60/FPA/(frames-5) Use the FPA listed above and remember that frames must be a multiple of 6 for Hydras (so 6, 12, 18, 24, 30, etc, etc). 30 frames would be 2 hydra attacks per second (60/30), 24 would be 2.5 hydra attacks per second (60/24), and so forth. The result of the formula would be the APS your wizard would need for that hydra attack speed (breakpoint). Plugging this forumula into Excel will give you an instant calculation of your hydra APS/BP if you like: =60/(ROUNDUP(ROUNDDOWN(60/FPAHydra/(APS*(1+BuffTaskers/100)),0)/6,0)*6) You'd just have to give it the FPA of the hydra you're using, your sheet APS and your Taskers% in the corresponding spots in the forumula.
  4. Looks solid. Get this SSD instead though https://www.newegg.com/Product/Product.aspx?Item=9SIA12K2GN1590 The WD SSDs are not noted as great performers. The Samsung is about as fast as it gets on SATA III. Price difference is $30 but I think it will be well worth it. You could also drop down to the 650W version of that PSU with no problems at all. It's still more than enough power for a single GTX 1060. https://www.newegg.com/Product/Product.aspx?item=N82E16817151158 Saves you $20 back from the SSD upgrade. That is a stellar PSU. You won't get much better than that, but you could save a few more dollars going with something not quite so top-of-the-line. You've got a 95W CPU and 120W GPU in there so your total draw isn't likely to be much over 250W-300W at peak (if even that). You could safely shop down into the 400-450W range if you wanted but the trade-off is that you're not going to get something engineered and built like a PRIME in that power range.
  5. That's in the notes. It's a separate record for each game type. So yes, no primals in a season until you've done GR70 in that season first.
  6. I was testing with MPC-HC 1.7.9 x64, which shouldn't make that much difference. It plays fine with hardware decode enabled, that wasn't the issue. I was more pointing out that software only is an issue for HQ 4k HEVC and that you do need it at the moment. 24 fps is not "low" for high quality. It's standard for films content (Blu-rays, etc). Bit rate is a far better indicator of quality. The overall bitrate of your sample was 54 MB, which is quite high and is ~4x or higher the rate of a most decent 1080p 24fps x264 stuff you typically find. Given that HEVC is also ~x2 as efficient on bitrate for parity quality of x264, that would work out to ~27 Mbps 1080p 24fps x264 and easily on par with blu-ray video quality. http://demo-uhd3d.com/fiche.php?cat=uhd&id=143 is the actual clip I was testing with. It's 60fps HEVC with 75 Mbps or so. Maybe it's entirely frame rate dependent but it was unplayable with software decode only. Incidentally, I can play that sample you linked without hardware decode at 40-50% CPU usage on my 4790k. Enabling hardware decode through dvxa (copy-back) works and knocks it right down to 3-6%. CUVID/CUDA hardware decode does not seem to work with that sample and crashes my MPC-HC. Maybe that's a 1.7.9 issue though.
  7. Mine is a 3930K as well. Figured we were probably talking about the same CPU. It was the most reasonable of the first gen socket 2011s. No. I could of course but the quality tends not to be as good as a straight software transcode. I admit I haven't revisited the problem in a while but there was a pretty wide consensus that the CUDA based encoders weren't as clean and accurate as x264. QuickSync wasn't too bad though but I don't bother with it. A typical transcode takes only 2-3 hours depending on the size and I generally run them in batches overnight if possible. What little QuickSync might cut from that isn't really going to make a lot of difference in my usage and I just stick with the best quality of software x264 instead. Decode and playback is a different matter though. You're not going to be doing software only playback of HQ 4k HEVC on any quad core, especially if it happens to be 60 fps. I doubt if an 8c would make any difference either. CUDA/CUVID makes a huge difference there and looks just fine. Software HEVC in 1080p is just fine but jumping to 4k seems to bring a massive performance hit that can't currently be overcome without a hardware decoder of some sort. Not really a big deal for me right now since I don't have any 4k displays and I'm not collecting any content yet either. I did recently do a test playback while trying to help someone troubleshoot some codec issues though and I was shocked at just how hard my i7-4790K choked on the sample file. It might also mean that transcoding 4k (if that ever even becomes viable) might require a hardware solution too.
  8. 'grats Machado. Sounds a lot like my old Sandy Bridge-E that I retasked into my office/x264 transcode workstation. Except 32 GB of RAM seemed very excessive 5 years ago so it's only 16GB. Yes. There is something that is just innately pleasing about seeing 12 CPU graphs in Task Manager. 16 would be even better. Unfortunately, aside from some transcoding (which I do on the 6c/12t anyways) it's pretty rare that I really max out my current 4c/8t in general usage for any significant amount of time on my primary gaming desktop.
  9. No argument here. I can also see why someone might consider an X1700 for a general desktop too. It depends a lot on exactly what you use it for. If you do mostly web browsing and office type stuff along with some moderate gaming (particularly at 1080p) where your GPU is already easily up to the task, then an Intel i5 probably makes more sense (unless you also keep a shit-ton of stuff open all the time at once that will actually use up all 4 cores). If you're running a 4k setup, then it's less of an issue (although eventually, as GPUs get upgraded you'd end up in the same boat). AMD's usage case for streamers also makes some sense, but that is definitely sighting a multicore friendly usage case. It looks like their memory controller is also weaker, so that could also be a factor in what solution is best for some people. Also, if games start to actually use 4+ cores on a regular basis, then Ryzen is a good choice. We've been talking about that for years though and it has yet to really happen. It could, but it's a usage scenario that doesn't really exist right now aside from a few rare exceptions. Ashes of the Singularity is one but also clearly shows signs of not playing well with Ryzen (for now at least) anyways. The other? I don't think I could justify slowing everything else down for that one game. It's been a long time since I used an AMD platform but I wasn't really happy with it back then... All signs do point to that having improved a lot as well so it's less of an issue. I could see myself seriously looking at a Ryzen if and when I decide to rebuild my AD fileserver. The old Q6600 is starting to struggle under Server 2008 with DFS active and more cores would probably work out well there. I'd have to look at what Xeon's are available but generally speaking it looks like Ryzen would be a far cheaper and better solution for that - and Intel's desktop 8c has priced itself out of contention right now. In Garble's case, it sounds like he does mostly the typical office/student workload type stuff. He also plays D3, which is not particularly GPU demanding and would probably benefit more from a strong i5/i7 than 6c/8c. I don't know what else he plays though. He may also choose to bank on the idea that multi-threading (using more than 4c effectively) is going to become much more prevalent. All in all, it's good that they're relevant enough that there are things you can weigh in making a decision. It will also be interesting to see what Intel's response is. Not just in pricing but in whatever hardware they might release in response over the next year or two.
  10. The Ryzen benchmarks this morning are showing more or less what I expected. "Big" Ryzen (the 1800X) competes very well with Intel's workstation series (the socket 2011 -E series, particularly on performance vs cost) but is showing ~25% slower than Intel's top mainstream 4c offerings like the 7700k (which is $150 cheaper). Gaming performance depends very much on how multi-threaded the game in question is and how GPU bound it is. There is also the question of non-optimized code paths yet too. Still... unless the 4c Ryzen CPUs come with a significant bump in clockspeed to put them more in parity with Intel's mainstream series, they're still behind. If not, then they had better be at least 25% cheaper than compartive Intel 2c and 4c offerings but even then, that puts them in the "budget" choice rather than the driver's seat. If that's the case, let's hope it's enough to keep the company healthy. The vast majority of the CPU market is made up of the mainstream/business and gaming market, not the workstation end. Overall, Ryzen is good and an impressive piece of hardware from AMD. I'm not sure they targetted the right segments but they do very well in what they did target (strong total multi-threaded performance). To me, it appears that Intel still has the better approach to the mainstream markets (despite there irritating segmentation of features). If I was building a new encoding station, an 8c/16t Ryzen would probably be the top contender. For a new desktop/gaming PC? Perhaps not so much, but we haven't seen the lower end of the lines play out yet either.
  11. The problem is that beyond basic OS multicore requirements, single threaded performance is still largely king in the mass market. A quad core (some would even argue a dual core) with higher per core performance is still the best solution for most people, including gamers. Games are starting to come around a little bit but there are very few titles that actually benefit from having 8 cores instead of 4 faster ones. Star Ruler 2 is the only game I have that actually makes be rethink the decision I made last upgrade to move from a hexcore back to a quad - everything else clearly runs better on the quad. Business/office stuff has been moving to laptops for a while now which is indication that max cpu power isn't the primary factor anymore, even Intel is promoting its NUC platform as the future of business PCs. Ryzen exceeding its stated IPC gains should help but I've seen nothing that indicates that it's going to match Kaby Lake's ST performance. So yes, it's going to put some pressure on Intel pricing but only performance-lessened value orientation way since for most things that actually drive the majority of PC sales and upgrades Intel will still hold the crown. They're always going to be able to justify some premium if the pre-review indications hold true. I don't expect to see Intel's Quads come down too much more in price, although I would like to see a sub $200 model. It would also be nice to see less pissing around with K vs non K and what has hyperthreading and doesn't unless it's i7 and "premium". The 6 and 8 core models though? I sincerely hope that Ryzen gives Intel a run for it's money at the prices they are talking, because in my opinion most of Intel's -Enthusiast line is priced twice as high as it should be (or more) and with motherboards priced in-step as well. Even then though, those 6+ core CPUs are only a big draw to the encoding and heavy workstation usage crowd. I don't think Ryzen has it made just yet and it won't surprise me if AMD still has a bit of an uphill battle to fight in the long term but this does seem like a solid step forward in getting back into the war. An Intel i5 quad with hyperthreading sitting in the 3.5Ghz with 4Ghz Turbo range priced at $250-$300 CAD (convert to USD as you will) would have a good chance to become an obvious go-to choice for most of the still performance vs value oriented crowd and would probably eat Ryzen's lunch. It should become more clear in a few days. Kaby Lake introduced an i3 with HT so it's definitely not out of the question.
  12. When an 8 core Intel is still $1,400 (CAD, admittedly) it's still hard to get too excited. The 6950 is still something like $2,200 if I wanted to get one. It's a start though. I'll be interested to see some thorough reviews done by the major tech sites. If Ryzen is truly competitive (the hype is certainly in full swing), then I'd hope that it's just the start of a trend.
  13. Thanks guys. Happy birthday Booz.
  14. Good. It's about time they did something about that skill. It made no sense that it cost more, was harder to sustain, but did less damage. It's still not going to do much good though (as a primary build damage source) without some support items to boost it.
  15. Top left? That's an HD-DVD player. Top Right is the blu-ray player. Middle right is an old DVD/SACD player - used for CD/SACD only. I want a better player but it's hard justifying the price for the usage.