I’ve never owned an Intel before, but next month I’ll buy one. I think I’ll go for the E6600 if nothing change until then… Best price/performance at the moment… http://www.tomshardware.com/2006/07/14/cor…_64/page12.html
Even the flagship from AMD (FX-62) is slower than the cheep E6600, so it will probably not be an AMD this time if not a miracle happens before the end of February…
And by the way, I really think multiple core support shall be at the most top of the prio list for renoise… Almost all new computer are multiple cores nowadays and if you buy one you can’t really take advantage of it while using renoise, and that would be a shame… so I really hope it wont take years trackit mentioned, but more like the first update we see after 1.8 is release, maybe within a couple of months or so…
I’m also about an upgrade, but really found E6300 is the best price/performance. The CPU can be overclocked to E6600 and above. It’s not that I’m going to overclock it, but E6600 is overpriced to me. FX-62 is a good CPU, but it’s just damn expensive for 90nm technology, just unreal expensive.
I just got back from our first tour where we used renoise live on stage. We have run into a lot of problems with renoise cutting out with certain songs, even if the CPU load isn’t high. It usually happens when there is a new pattern which introduces new samples (longer than one-hits). it seems inconsistent, and we are pretty good about removing unnecessary processes from the task manager before going on. but, it really, REALLY sucks when you are running a track live through a huge PA system and the song cuts out because renoise decides to choke at %60 cpu.
we found a weird work around to this issue… sometimes, the song will cut out only when it is played the first time through. if you let the track play through the “problem areas”, it seems to cache, or buffer the samples in ram or something, so if you hit stop and then start from the beginning, it will play through just fine. but it doesn’t make sense because isn’t the song already loaded all the way into memory? I don’t know if this is a bug or if we are just being too demanding… but i wanted to let you guys know about our problem.
Regardless of all of that, what are the most important features for a computer running renoise? Ram? Ram speed? CPU speed? Cache? hard disk speed? does dual core have ANY advantage at all? Doesn’t the operating system use one core for system processes while renoise might use another? I am going to buy a new laptop soon, and THE MOST important thing for me is that it has the hardware features renoise needs the most in order to play live tracks without cutting out unexpectedly.
p.s. i use a motu 828 firewire interface with a relatively high latency (1024 buffer).
Renoise loads all samples into memory, but as soon as the physical memory is full, or the system decides to swap it out to disk (when minimizing the application for example), it will take some time to get them swapped back into memory. This “swapping back” costs a lot of time, so you might get dropouts. Your workaround made sure that all samples (which are currently needed) are in memory (by simply using/playing them back).
You can do a few things to make that less likely:
get more RAM (so the chance that something gets swapped out is less likely)
get a faster HD (so when there something gets swapped, it will at least be faster)
defrag your HD (make especially sure that the pagefile is somewhere in the earlier sektors of your HD and and not fragmented)
use less sample memory (remove unused samples)
avoid minimizing the application during the gig
We should sooner or later (more sooner) introduce direct from disk reading for large samples. This will of course also make this happen less often, as less memory is used. For now you could use (as a workaround) some external VSTs to play back those really really large samples.
I don’t think the question is dumb, a very valid point imo
I have found myself getting a little frustrated with my curent work in progress. I’m running Renoise on a pretty heavily overclocked Opteron 170 right now (as in signature), and like to run at a low enough latency so anything I do during live sets isn’t too affected, but there is also a point where latency makes little difference to CPU usage, and I seem to be coming closer to my current setup’s limits. I know a lot of people will say render to samples, and lower samplerate, but I am always tweaking the tunes, and am never settled with them, so I like to keep the original elements of it as much as possible without loosing quality, or flexibility, especially during sets.
So anyway, my point being - there’s another core, doing next-to-nothing, but that is my only niggle atm
Luckily, I’m moving over to my new rig, an E6300 & Asus P5B-Deluxe. This was frighteningly faster than my overclocked Opteron at stock speeds, and am currently testing at 2.8GHz on just a trickle over 1 volt, perhaps it will go lower!? This chip seems to also hit a 100% overclock quite happily, but my memory is my limiting factor at the moment.
Sorry, I’m just bragging now, but the potential of C2D must surely be realised
Of course we are working on this. We would be stupid if we would waste half or more of the power of nowadays computers. Just let us finish the release first, then come up with new stuff. Releasing this monster is really hard enough for now…
I’ve been using renoise on a Core Duo T2300 lappy for last 6-7 months.
its been superb , considering I only have 512MB ram and onboard sound card. Using manymany vsti,vst and effects at once.
The speed of the computer and the fact some OS tasks are propbably taken on by least used core makes a difference … I have noticed a HUGE inprovement from my s478 3.0E
Sorry Taktik, I’m aware you & the team will be slaving away for us on this, and I didn’t mean to aim that at you guys, (you are all of course, doing a superb, and very attentive job). To be honest, my quarms are probably down to me being inefficient with my tracking anyway!
I was just trying to make the point that C2Ds are unbeatable for their price/performance
thanks! this is good info. though we never minimize renoise during the set. the samples we were playing in the track weren’t terribly huge. i think in some cases they might might have been a rendered channel from one pattern, or actually now that i think about it, we do sometimes use these multi-sample instruments for horns or violins which have those big professional orchestral samples. anyway thanks again.
Also (we didn’t do this live, but we will eventually) we have been using the clock sync feature, so that one computer controls the other. there is some “straying” from perfect sync and it is noticeable at times, but it always manages to find it’s way back after a dozen ticks or so.
the new version of renoise is great, btw, thanks for all the hard work. we appreciate it very much and will continue to use it for all of our composing and live sets.
Anyway, so what I’m really curious about now is, will a E6600 clocked at 2400 and not using it’s 2 cores potentional to the fullest with Renoise, outperform my P4 without hyper threading clocked at 3200?
And is it worth the investment?.. I really want more CPU power for Renoise…
yeah that’s my problem, too. i’ve got P4 3.4@3.9GHz and still, it’s not enough (sometimes =). i’ve got many expectations in dual core processors but i won’t buy it now because i’m afraid it would handle even less VSTs… yet…
@nula
the pentium4’s rely on intel’s old and dusty netburst architecture with their long execution pipelines. they have a very inefficient instructions-per-clock ratio, especially compared to core2duos.
i switched from a p4 3.0@3.8 to an e6600 @2x 3.5ghz and it’s about three times as fast in renoise now, even though one of the cores is not being used for audio prcessing.
even at stock clocks, any core2duo would outperform your 3.9ghz p4 power plant.
neither nor.
pentiums are all outdated and the product name will die a slow marketing death.
athlons are inferiour to intel’s current line of core2duo CPUs.
so, as mentoned above, the Intel Core2Duo CPU is currently the best performing processor to get for any kind of application.
AMD’s available CPUs are all out of competition at the moment.
every operation a CPU has to execute is mathematical.
however athlons do feature more processing power for floating point operations than pentium4’s. but then again, core2duo’s feature more power on the FPU side than athlons do.
at the very moment it has never been so simple before to decide about what CPU to buy.
maybe AMD’s upcoming “barcelona” K10-Athlon will be able to bring back the performance crown again… but it’s still a long way until its release and there are no official benchmarks available yet for that quad-core CPU.
Remember that more power doesn’t mean better music, or even ‘good’ music + Don’t get to hung up on upgrading technical specs and the utopian quest for the ultimate computer, you can lose a good deal of yer life on that kind of shit while you’re better of tracking some music!
that’s almost a philosophical question you arouse here.
of course you can write great music on c-64 or with 96k of samples, but the more available resources you got, the less limited you are in what you want to achive musically.
if i was still writing stuff on an Pentium3, i’d probably end up freezing and bouncing tracks all day long as soon as i exceed 4 times polyphony with my favourite CPU hungry software synth.
workflow would suffer greatly, even though i could actually write the same song as i could on a faster machine - but with a way greater effort.
i am as most humans: lazy, preferring the easier way.
and as long as it is feasible, i choose “realtime” over “post-processed”.
Damnit keith, you got me into thinking of buying a faster PC again. On the other hand, I keep my composing more minimal now, knowing I don’t have unlimited CPU-power, so I guess something good comes out of it, in the end.
Still, one day I need to upgrade my lovely machine.
