Separate names with a comma.
Discussion in 'TiVo Premiere DVRs' started by logicology, Oct 10, 2011.
I dont think there is any reason to doubt the Non-Elites wont see this software
I'm not unconfused by the non-singular instances of negatives in this statement.
For all we know, it could be a hardware fault that prevents the second core from working in the Premiere.
In another recent thread a Premiere user showed screen shots of the streaming capability using SW 14.8c.
I'm still trying to figure out what it means...someone, in English please
He is also the only person who can do that, none of the other Elite users have been able to successfully do what he's doing.
then if it's a hardware fault, is Tivo going to send all of us "beta-testers" units with working dual cores that we actually paid for? The sales pitch they gave me on the phone was that the dual core functionality was weeks away and depended on a software tweek.
That was in April, 2010.
I'm not being negative, but what would Brian Boitano do?
That can still be measured in weeks. It seems to me you have put too much credence into what you heard from a CSR a long time ago. They are not bad people, but many times they don't understand what they are being told, or the techies tell them things which eventually don't pan out. Sounds like the latter.
Me, I'll be happy to see whatever performance improvements can come from 14.9. Anyone laying the groundwork for this to be a hardware problem is simply being negative to a degree that is not necessary, as we simply do not know.
After looking over these threads, my glass is half full.
These chips are supplied by Broadcom, these aren't unique to the TiVo so other devices have them too. If there was an issue in the hardware it would have been resolved long ago. This realistically has been a software/programming issue that has gone unresolved until recently. Multithreading is not an easy feature to just turn on and make work. From the Impressions thread we can see they've moved up several versions on the GCC compiler as well as a newer build of the Flash Lite runtime layer as well as a new HDUI itself. Who knows what combination of changes finally made the 2nd code usable.
If a Tivo rep, REGARDLESS of where in the company they are sitting, sells the product on false-pretenses, then Tivo is responsible. It's not "laying groundwork", and it certainly isn't "negative", it's just business. They made a claim - they need to fulfill it.
And after waiting for 18 months, I'm ready for satisfaction, and a FULL glass.
My point is exactly that, we do not know whether it's a hardware problem or a software problem. It's unrealistic to speculate that 14.9 will add dual core support to non-elite Premieres. Tivo has given us no indication that it will, as they don't value the user community enough to share their plans.
I certainly hope it is a software issue and will be fixed with 14.9, but this company as of the last few years has taught me to assume nothing about the product.
Multithreading is a separate issue from SMP. SMP can be turned on in the Linux kernel with a simple recompile. Multithreading may require rewriting portions of the application code. A single-threaded app can be run in a SMP environment. One would expect at the very least we'd have an SMP kernel up by now, possibly running a single-threaded flash environment (assuming that popular speculation is correct, and it's a flash problem).
This isn't rocket science, it's not cutting edge research, it is old technology that's been in use for decades. There's no excuse why it should take 18 months to get it working. There's also no good reason why it should be rolled out on a new product rather than fixing the current product.
It is not just that TiVo doesn't share their plans, they often/usually don't even share information AFTER they make changes. They never post a list of what bugs are fixed or any lower-level information about updates even AFTER they are released. I guess that would be memorializing the fact that there were bugs, to begin with, in writing. Heaven forbid...
Since the Elite can easily handle reading/writing 7 concurrent HD streams(with the HDUI still faster than the two tuner models) they would have had no choice since it needs the extra performance to accomplish this. The Premeire can't handle that many without having issues. So since it isn't enabled in the two tuner models they could wait on it. But with the Elite, there is no way the Elite would work right running what the two tuner models are running.
How do you KNOW the NE Premiere can't handle recording 4 while playing back one with single core? And with or without the newer firmware? I don't think that is something we can test...
Again, we don't know how much performance improvements come from other parts of the firmware/OS/Flash/UI being changed or updated...
Even though the Tivo Premiere/Elite uses a different Flash architecture, multithread support is coming to Flash.
How do you have one thread executed by more than one processor? It might not necessarily be symmetric, but surely use of multiple processors implies use of multiple threads. If SMP is simply turned on in the Linux kernel, then at least the kernel would be multithreaded. Of course, it's the application code that we want to see multithreaded, and Flash.
Given the fact that multithreading has already been delayed for so long and that there is widespread skepticism about TiVo's ability to deliver it, they would be crazy to roll it out to all Series 4 units when they can instead test it in a smaller population of new units which people will expect to have some teething problems anyway.
I'm going by when they had steaming enabled. I got 5 or 6 HD streams reading/writing, it worked but it did have issues. And one of the streams, the download was going slow, but it was doing it.
It depends on what you mean by "use" and in what context. To utilize SMP requires either multithreading or multiprocessing. I could run a single Flash app in one core, and a different single-threaded app in the other core. The kernel can be background flushing data in one core while the other core is doing computation. I'm assuming the Tivo already makes use of multiple processes, that the Flash GUI is a separate process than the backend core that is handling tasks like recording and scheduling and applying database updates. SQL queries and such are probably executed in a separate process. Even on a single core, these multiple processes already have context switching and synchronization in place. There's already parallelism.
You're right though, if the primary consumer of CPU is just one app, and that app is single-threaded, then one core is going to be idle most of the time.
Perhaps I'm old-fashioned, but I think my purchases should be tested and functioning before they're sold to me. That includes the Premiere Elite. If dual core is not ready for prime time, then it shouldn't be rolled out on any DVRs other than people participating in the beta program. Consumers purchasing Elites are not beta testers. Shipping an untested Premiere is how we got into this mess in the first place; shipping an untested Premiere Elite is not the solution.
The poster said "I dont think there is any reason to doubt the Non-Elites wont see this software"
If I were to replace "I dont think there is any reason to doubt" with "I'm sure that", then we get "I'm sure that the Non-Elites wont see this software"
Is that what he meant?
But since that is streaming, and not "recording" or "playing back", you might have hit maximum network overhead. It is not uncommon for 1080i CATV HD streams to be between 15 and 17 megabits. Since we have proven the Premiere network is around 85-90 megabits max, that would be on 4 to 5 streams before saturating the network. If you had any OTA 1080i content, that could climb over 22 megabits for just one stream (rare, but possible). It is no big deal with local "recording" or "playback", however, since that is to the hard drive, which is much, much, much faster.
Plus, again, it is a different version of all the firmware (Linux/Flash/UI/etc), which really could make a big difference.