Range 1, 445T-450T done, 12480 factors: [URL]http://uwin.mine.nu/TPS/480000-485000/factorupload/tpfactors_n480K-485K_p445T-450T.txt.bz2[/URL]

Reserving 460T-465T.

450T-455T for n=480K-485K complete, 12458 factors.

[url]http://www.sendspace.com/file/lsghkv[/url]

Reserving 450T-455T for n=485K-490K

[QUOTE=Robert_47;222744]Reserving 450T-455T for n=485K-490K[/QUOTE]

Did you mean 465T-470T?

[quote=gribozavr;222747]Did you mean 465T-470T?[/quote]

No, I was thinking about the second range, for 485000<n<490000. Unless you want it, I'm not picky.

Well, I'm sieving n=485000-495000 together. So if you haven't yet started, please take a range from n=480000-485000.

I take 460T-550T for n=485000-495000.

[quote=gribozavr;222776]Well, I'm sieving n=485000-495000 together.[/quote]

That's what I get for not reading the old reservation posts.:smile:

I haven't started yet, so I'll release 450T-455T for n=485K-490K.

I've updated the reservation and stats pages to try and sort out the confusion. Here's what happened:

- gribozavr's 485K-495K reservation was extended from 460T-550T to 450T-550T. That's because nobody was working on the 450T-460T range for 485K-495K.

- Robert_47 was assigned the 465T-470T range for 480K-485K.

If either of you object to those arrangements, feel free to post. Otherwise, the last few posts will be moved to the "Archived sieve reservations and discussions" thread in 2-3 days.

[QUOTE=Oddball;222784]- gribozavr's 485K-495K reservation was extended from 460T-550T to 450T-550T. That's because nobody was working on the 450T-460T range for 485K-495K.[/QUOTE]

I'm fine with this.

It's fine with me. Sorry for the confusion.

Range 1, 460T-465T done, 12141 factors: [url]http://www.sendspace.com/file/hdg57b[/url] (Lennart's upload site is down)

Reserving 470T-475T.

mdettweiler, Robert_47: I've downloaded and verified your files.

465T-470T complete, 12283 factors.

[URL]http://www.sendspace.com/file/cyv2r5[/URL]

[QUOTE=Oddball;220836]I just checked Joshua2's profile and found that he hasn't logged on since June 10th. I'm allowing him a few more weeks to account for things like long vacations and other delays, but if he doesn't respond by the end of the month, his 200T-300T reservation will be canceled. The range doesn't have to be finished by that time, but we need another status update to be sure it's still being worked on.
[/QUOTE]
Joshua2's reservation has been canceled, so 200T-300T is now available.

470T-475T (Range 1) complete, 12054 factors: [URL]http://uwin.mine.nu/TPS/480000-485000/factorupload/tpfactors_n480K-485K_p470T-475T.txt.bz2[/URL]

Reserving 475T-480T.

Robert_47, mdettweiler: I've downloaded and verified your files.

400T-550T for n=485000-495000 finished.

Taking 550T-750T for n=485000-495000.
Taking 200T-300T for n=480000-485000.

455T-457.5T complete, about 6100 factors found.

Part 1 (n=480000-482500): [url]http://www.sendspace.com/file/m1ewqx[/url]

Part 2 (n=482500-485000): [url]http://www.sendspace.com/file/h36fgr[/url]

There's a heat wave rolling through the area I live in, so I'll release 457.5T-460T. Temperatures should come down by the end of the week, so if nobody picks up the range before Friday, I'll finish it off then.

Taking 457.5T-458T, 480K-485K.

457.5T-458T complete, 1189 factors found:

[URL]http://www.sendspace.com/file/ovrk6c[/URL]

Taking 458T-460T.

[quote=MooMoo2;223630]Taking 457.5T-458T, 480K-485K.[/quote]
Welcome! It's nice to get more computing power on here.

Oddball, MooMoo2: I've downloaded and verified your files.

I will be out of town until 14th so I won't able to download results until then. Fortunately, sendspace keeps files for 30 days.

[quote=gribozavr;223802]
I will be out of town until 14th so I won't able to download results until then. Fortunately, sendspace keeps files for 30 days.[/quote]
I'll be away too for a while. My vacation's a bit shorter - from August 7th until the 12th.

Is now Moomoo2=MooooMoo ?

475T-480T (Range 1) done, 11728 factors: [URL]http://uwin.mine.nu/TPS/480000-485000/factorupload/tpfactors_n480K-485K_p475T-480T.txt.bz2[/URL]

Reserving 480T-485T.

458T-460T complete, about 4760 factors (don't remember the exact number):

[URL]http://www.sendspace.com/file/9s1dtx[/URL]

I'll pick up another range when I come back from vacation late next week.

[quote=Oddball;223885]I'll be away too for a while. My vacation's a bit shorter - from August 7th until the 12th.[/quote]
My new computer allowed me to squeeze in one more completed range before I go on vacation.

485T-490T is complete, and about 11400 factors were found:

[URL]http://www.sendspace.com/file/g0sj2g[/URL]

edit: gribozavr, could you post a new file once you're done with the 200T-300T range? The links were posted about a month and a half ago, so some of the sendspace links have expired. Also, it'll be nice to see how many k's are remaining now.

[quote]My new computer allowed me to squeeze in one more completed range before I go on vacation.
[/quote]
Well, I got to squeeze in yet another range. This computer rocks!
490T-495T complete, about 11200 factors found:

[URL]http://www.sendspace.com/file/dwsnad[/URL]

My flight leaves in just a couple of hours, so this will really be the last range I finish before going on vacation.

480T-485T (range 1) complete, 11592 factors: [url]http://www.sendspace.com/file/vche0n[/url] (Lennart's site is acting a little strangely ATM.)

Reserving 495T-500T.

BTW @Oddball: I'm not sure if you'll get this before departing, but did you by chance run a comparison to see which is faster overall, one 6-threaded instance of tpsieve or 6 one-threaded instances? In my experience, even the best of multithreaded applications suffers a little performance hit as it scales across multiple cores; I'm curious to see how pronounced that effect is with 6 threads.

[quote=mdettweiler;224391]480T-485T (range 1) complete, 11592 factors: [URL]http://www.sendspace.com/file/vche0n[/URL] (Lennart's site is acting a little strangely ATM.)

Reserving 495T-500T.

BTW @Oddball: I'm not sure if you'll get this before departing, but did you by chance run a comparison to see which is faster overall, one 6-threaded instance of tpsieve or 6 one-threaded instances? In my experience, even the best of multithreaded applications suffers a little performance hit as it scales across multiple cores; I'm curious to see how pronounced that effect is with 6 threads.[/quote]
From memory tpsieve is a memory hog. He probably doesn't have enough memory to run 6x1Gb jobs

[quote=henryzz;224451]From memory tpsieve is a memory hog. He probably doesn't have enough memory to run 6x1Gb jobs[/quote]
Hmm, hadn't thought of that. I suppose 2 instances with 3 threads apiece would work pretty well, though--depending on how tpsieve scales (the only other comparison point I have is Prime95, which does a radically different type of calculation; it scales badly for >4 threads), that may do the trick to recover a lot of performance loss.

When I was sieving for n=333333 I got the best performance by running -t2 sieves which was as fast as two single threads but less administration. Two times -t2 on a quad, four times on a V8. More threads per sieve were inefficient even when I did not reach memory limits.

BTW: which is the most recent release for Linux64? Maby newer version behave differently...

495T-500T complete, 11300 factors: [url]http://www.sendspace.com/file/1mgk1w[/url]

Will return after the impending NPLB rally. :smile:

[quote=Oddball;223885]I'll be away too for a while. My vacation's a bit shorter - from August 7th until the 12th.[/quote]
I'm back :smile:
Reserving 500T-510T.

[quote]
Is now Moomoo2=MooooMoo ?[/quote]
Probably not, the name seems to be a Pokemon reference and nothing more. In that game, MooMoo milk was an item used to heal Pokemon. It came from a Pokemon called Miltank, which also happens to be MooMoo2's avatar. As for the number 2 in MooMoo2's username, I'm guessing that it's there because Miltank first appeared in the 2nd generation of Pokemon.

Anyway, I'm not merging the MooMoo2 and the MooooMoo names in the stats page. MooMoo2 =/= MooooMoo unless proven otherwise.

[quote]
I'm not sure if you'll get this before departing, but did you by chance run a comparison to see which is faster overall, one 6-threaded instance of tpsieve or 6 one-threaded instances? In my experience, even the best of multithreaded applications suffers a little performance hit as it scales across multiple cores; I'm curious to see how pronounced that effect is with 6 threads.[/quote]
[quote]From memory tpsieve is a memory hog. He probably doesn't have enough memory to run 6x1Gb jobs[/quote]
I have 8GB of RAM, which is barely enough to fit 6 X 1GB threads (Windows 7 takes up nearly one and a half gigs of RAM). Here were the results:

one 2-threaded instance of tpsieve: 33.1 M p/sec
two 1-threaded instances of tpsieve: 33.3 M p/sec

one 6-threaded instance of tpsieve: 85.2 M p/sec
six 1-threaded instances of tpsieve: 84.9 M p/sec

The performance hit/gain seems to be negligible for tpsieve.

[quote]
495T-500T complete[/quote]
That finishes a major milestone - the work done since last year is equivalent to sieving n=480K-495K until p=500T (the 480K-485K range is missing 200T-300T, but this is balanced out by the 485K-495K range, which is completed to 550T). Let's see whether we can reach 1P by the end of next month.

Thanks to everyone who participated!

Reserving 510T-515T (range 1).

500T-510T complete, about 22100 factors found:

[URL]http://www.sendspace.com/file/bh3z6d[/URL]

Reserving 515T-525T.

I'm back. I've downloaded and verified all files from posts above.

Oddball: Unfortunately, it was (is?) extremely hot here and we had problems with electricity. Range 200T-300T should have been completed, but it is not, I'm now continuing it (32% done). The machine that was processing 550T-750T is a remote one and it is still down. I don't know how much of 550T-750T is done, so in order not to duplicate work, I'll be waiting until the machine is fixed.

515T-525T complete, about 21300 factors found:

[URL]http://www.sendspace.com/file/8mbks2[/URL]

I'll get 525T-550T.

[quote=gribozavr;225500]
Oddball: Unfortunately, it was (is?) extremely hot here and we had problems with electricity. Range 200T-300T should have been completed, but it is not, I'm now continuing it (32% done). The machine that was processing 550T-750T is a remote one and it is still down. I don't know how much of 550T-750T is done, so in order not to duplicate work, I'll be waiting until the machine is fixed.[/quote]
I see. When you're done with the 200T-300T range, could you upload the 480K-485K file as well as the small k file (k<100000 for n=480K-481K)? I might try some LLR work for a change.

Reserving 550T-555T.

For benchmark purposes, I've completed 5999T-6000T.

171 factors were found. The file is quite small, so I've attached it to this post instead of uploading it to sendspace.

edit: After finishing this range, I've found that 6P is the optimal sieve depth. See the "Two Sieving Questions" thread for more details.

510T-515T complete.

[url]http://www.sendspace.com/file/1lhgv4[/url]

Taking 555T-565T.

550T-555T finished, 10220 factors found:

[URL]http://www.sendspace.com/file/1amwm1[/URL]

Reserving 565T-570T.

525T-550T complete, about 51700 factors found:

[URL]http://www.sendspace.com/file/ra2laz[/URL]

Reserving 570T-585T.

Reserving 585T-590T.

I'm done with 565T-570T. 9904 factors were found:

[URL]http://www.sendspace.com/file/kd4zcg[/URL]

570T-585T complete, about 28900 factors found:

[URL]http://www.sendspace.com/file/b1wiqc[/URL]

Reserving 590T-600T. There's yet another heat wave rolling through my area, so I'm reserving a smaller range this time.

555T-565T complete, about 19800 factors.

[url]http://www.sendspace.com/file/c6198j[/url]

585T-590T done, 9481 factors: [URL]http://www.sendspace.com/file/dn6ygd[/URL]

Reserving 600T-605T.

I've downloaded and verified all files.

200T-300T complete. Taking 605T-700T.

590T-600T complete, about 19100 factors found:

[URL]http://www.sendspace.com/file/siv23n[/URL]

gribozavr, could you upload the small k file (k<100000 for n=480K-481K)? I'm going to do some LLR tests for a change.

edit: as a reminder, please make sure that there are no factors < 600T for the small k file.

Reserving 700T-705T.

700T-705T complete, 7875 factors found:

[URL]http://www.sendspace.com/file/28mech[/URL]

600T-605T complete, 9279 factors: [URL]http://www.sendspace.com/file/5trh5z[/URL]

Resrerving 705T-710T.

480000-484999_09dec2010.zip [URL]http://www.sendspace.com/file/uy637c[/URL]

Stats (number of k's):
18013514 480000-484999_30aug2010.txt
16904938 480000-484999_09dec2010.txt
6.15% removed

Small k file for LLR:
n_481500-484899__k_1_100000__28nov2010.txt [URL]http://www.sendspace.com/file/ul26bb[/URL]

I've downloaded and verified all files. 480000-484999_30aug2010.txt and n_480000-480999__k_1_100000__30aug2010.txt have all those factors removed.

705T-710T complete, 7954 factors: [URL]http://www.sendspace.com/file/4c5q6w[/URL]

Reserving 710T-715T.

I was originally planning to do some LLR, but seeing Oddball test thousands of candidates without finding any primes has been truly discouraging. Oh well, I guess I'll sieve instead.

Reserving 715T-720T.

[quote=MooMoo2;228193]I was originally planning to do some LLR, but seeing Oddball test thousands of candidates without finding any primes has been truly discouraging. Oh well, I guess I'll sieve instead.[/quote]
At this sieve depth (605T contiguous) and n- and k-level, the odds of finding a (non-twin) prime on a given candidate are 1 in 5490--that is, one would expect a prime about every 5490 candidates. The odds of finding one prime in the n=480071-480200 range currently loaded into the PRPnet server (4645 candidates) are about 57%. So it's a little better than a toss-up that there will be at least one prime found within the next few days at current processing rates.

How many cores do you have? Oddball's got 6, but you can still crank out a decent amount of pairs (enough to likely get a prime within a couple of weeks) with 4 or even 2 cores. With 1 core it would take a bit longer, but a prime is still not attainable.

Note that despite all my above words to encourage you that finding a prime isn't really impossible :smile:, it actually is better in the long run to sieve instead at this time. We're still quite a ways from optimal depth at p=6P, so running LLR tests right now is not really the most efficient way to go. (It's more to break the monotony of sieving than anything else. :smile:)

[QUOTE=mdettweiler;228202]With 1 core it would take a bit longer, but a prime is still not attainable.[/QUOTE]

A "not" seems to have sneaked into your post.

[quote=10metreh;228203]A "not" seems to have sneaked into your post.[/quote]
Ah, right--thanks for catching that. I meant to say "not entirely unattainable". :smile:

[quote=mdettweiler;228202]At this sieve depth (605T contiguous) and n- and k-level, the odds of finding a (non-twin) prime on a given candidate are 1 in 5490--that is, one would expect a prime about every 5490 candidates. The odds of finding one prime in the n=480071-480200 range currently loaded into the PRPnet server (4645 candidates) are about 57%. So it's a little better than a toss-up that there will be at least one prime found within the next few days at current processing rates.[/quote]
There's already been a lot of LLRing going on. You have about 3050 tests from the manual reservations, 2212 from PRPnet, and another ~500 from LLRnet. That's over 5700 tests, and it doesn't include the ~10,000 tests done in the n=390,000 subproject. So TPS has gone primeless despite testing about 16,000 k/n pairs.

BTW, what are the chances of finding a twin in the 480000<n<485000 range for k<10M?

[quote]
How many cores do you have? Oddball's got 6, but you can still crank out a decent amount of pairs (enough to likely get a prime within a couple of weeks) with 4 or even 2 cores. With 1 core it would take a bit longer, but a prime is still not attainable.[/quote]
I'm putting 1-2 cores on this project (the rest of my cores are working on an individual k reservation).

[quote]
Note that despite all my above words to encourage you that finding a prime isn't really impossible :smile:, it actually is better in the long run to sieve instead at this time. We're still quite a ways from optimal depth at p=6P, so running LLR tests right now is not really the most efficient way to go. (It's more to break the monotony of sieving than anything else. :smile:)[/quote]
At p=6P, what are the odds that a k/n pair would be prime? Would it be much better than the 1 in 5490 odds of finding a prime at today's depth of p=605T? Just wondering.

[quote=MooMoo2;228206]There's already been a lot of LLRing going on. You have about 3050 tests from the manual reservations, 2212 from PRPnet, and another ~500 from LLRnet. That's over 5700 tests, and it doesn't include the ~10,000 tests done in the n=390,000 subproject. So TPS has gone primeless despite testing about 16,000 k/n pairs.[/quote]
Note that when I say we'd expect a prime every 5490 candidates, that actually means that after testing exactly 5490 candidates, there is a ~62% chance that a prime will have been found. (Why this is involves some more advanced probability math which I myself don't entirely understand; I picked up most of this stuff from Gary at the NPLB project who is our resident probability whiz.) So having tested ~5700 tests so far on the variable-n drive, there's a 64% chance that we'll have found a prime so far--or, a 36% chance that we [i]won't[/i] have found a prime yet. Thus, what we have observed is an event that is statistically predicted to happen 36% of the time--which is well within the realm of possibility.

Informally, I could say that we are "due" for a prime at this point. But that's techincally an expression of the "gambler's fallacy" since having not found a prime so far does not have any bearing on whether we will find one soon (because they are independent events). More precisely, I could say that I'd [i]expect[/i] to find a prime some time soon because as we test more pairs, it becomes less and less statistically likely that we won't have found a prime.

[quote]BTW, what are the chances of finding a twin in the 480000<n<485000 range for k<10M?[/quote]
For the entire n=480K-485K, k<10M range we would expect:
-3264.835 single primes
-0.592 twins

Put otherwise, there's a ~45% chance that we'll find a twin in this range. Multiplying that by 4 for the entire n=480K-500K range, we get a ~90% chance of finding a twin.

[quote]I'm putting 1-2 cores on this project (the rest of my cores are working on an individual k reservation).[/quote]
In that case, then, a prime should be quite reasonably attainable. I would expect two cores running 24/7 on LLRing to find a (non-twin) prime within a couple of weeks. (Very rough estimate, though it should be approximately correct.) Note, of course, that one can always "beat the odds" and still not find a prime after that long.

[quote]At p=6P, what are the odds that a k/n pair would be prime? Would it be much better than the 1 in 5490 odds of finding a prime at today's depth of p=605T? Just wondering.[/quote]
The odds that a particular k/n pair will be prime do not change with sieve depth. However, what does change is that a lot of the pairs are eliminated during sieving, so it doesn't take as long to search the same range.

[quote=mdettweiler;228212]Note that when I say we'd expect a prime every 5490 candidates, that actually means that after testing exactly 5490 candidates, there is a ~62% chance that a prime will have been found. (Why this is involves some more advanced probability math which I myself don't entirely understand; I picked up most of this stuff from Gary at the NPLB project who is our resident probability whiz.) So having tested ~5700 tests so far on the variable-n drive, there's a 64% chance that we'll have found a prime so far--or, a 36% chance that we [I]won't[/I] have found a prime yet. Thus, what we have observed is an event that is statistically predicted to happen 36% of the time--which is well within the realm of possibility.
...
For the entire n=480K-485K, k<10M range we would expect:
-3264.835 single primes
-0.592 twins

Put otherwise, there's a ~45% chance that we'll find a twin in this range. Multiplying that by 4 for the entire n=480K-500K range, we get a ~90% chance of finding a twin.
[/quote]
I see. Thanks for the explanation.

Oddball, you can move this discussion to a seperate thread if you want.

PRPnet and LLRnet are both down, so I'll take 720T-730T.

710T-715T complete, 7482 factors: [URL]http://www.sendspace.com/file/4frdyp[/URL]

Reserving 730T-735T.

Reserving 735T-740T. Lennart I will test this on CUDA. Only a slow 250 card.

605T-700T complete. Taking 740T-900T.

715T-720T complete, 7288 factors found:

[URL]http://www.sendspace.com/file/bs64ce[/URL]

[URL]http://pgllr.mine.nu/privat/local/tps_735T-740T.txt[/URL]

[URL="http://pgllr.mine.nu/privat/local/tps_735T-740T.txt"][/URL]735T-740T complete


Lennart

720T-730T complete, about 14500 factors found:

[URL]http://www.sendspace.com/file/cayaop[/URL]

[QUOTE=mdettweiler;228212]that actually means that after testing exactly 5490 candidates, there is a ~62% chance that a prime will have been found. (Why this is involves some more advanced probability math which I myself don't entirely understand;[/QUOTE]

If you ran many many tests you would expect to find one per 5490 tests. But in any particular set of tests you might, if you were lucky, find two or three or four or even more. The only way both statements can be true is that there must be some runs of 5490 that find none.

It's pretty basic probability that if the probability of finding none in one test is 5489/5490, then the probability of finding none in 5490 independent tests must be (5489/5490)^5490. Nothing advanced yet, and that is good enough to get your 62% figure.

The advanced part comes from the observation that this is approximately e^(-1), and a few other limit consequences.

William

730T-735T complete, 7102 factors: [URL]http://www.sendspace.com/file/lnqulo[/URL]

Reserving 900T-905T.

Reserving 905T-910T. Lennart

550T-750T for n=485000-495000 finished. Taking 750T-950T.

[URL]http://pgllr.mine.nu/privat/local/tps_905T-910T.txt[/URL]

905T-910T complete

Reserving 920T-925T


Lennart

910T-920T complete, about 11300 factors found:

[URL]http://www.sendspace.com/file/x6ly6k[/URL]

I'll get 925T-930T.

[URL="http://pgllr.mine.nu/privat/local/tps_920T-925T.txt"]http://pgllr.mine.nu/privat/local/tps_920T-925T.txt [/URL]

920T-925T Complete

Reserving 930T-940T

Lennart

925T-930T complete, about 5700 factors found:

[URL]http://www.sendspace.com/file/k8o6d0[/URL]

Reserving 940T-950T.

940T-950T done:

[URL]http://www.sendspace.com/file/a46j8s[/URL]

About 11100 factors found.

edit: Just realized that this is the 2500th post in the TPS forum.

[URL="http://pgllr.mine.nu/privat/local/tps_930T-940T.txt"]http://pgllr.mine.nu/privat/local/tps_930T-940T.txt [/URL]

930T-940T Complete Lennart

900T-905T complete, 5713 factors: [URL]http://www.sendspace.com/file/w449t1[/URL]

Reserving 950T-955T.

Reserving 955T-1000T.

I think it'll be great to reach the 1P milestone by the end of the month :smile:

955T-1000T complete, about 48300 factors found:

[URL]http://www.sendspace.com/file/xf3itp[/URL]

Reserving 1000T-1005T.

950T-955T done, 5466 factors: [URL]http://www.sendspace.com/file/ywewyx[/URL]

Reserving 1005T-1010T for myself; and 1010T-1015T to test Gary's new GPU (this one should be marked under user gd_barnes).

1010T-1015T complete, 5192 factors: [url]http://www.sendspace.com/file/s6j056[/url]

Just as a reminder, the points for this should go under gd_barnes, not me.

I've downloaded and verified all files.

1000T-1005T complete, 5307 factors found:

[URL]http://www.sendspace.com/file/lgxvs5[/URL]

1015T-1025T complete, about 10300 factors found:

[URL]http://www.sendspace.com/file/2mw4ib[/URL]

1005T-1010T complete, 5195 factors: [URL]http://www.sendspace.com/file/hpncyo[/URL]

Reserving 1025T-1030T.

Reserving 2000T-2010T (just wanted to check the factor density at that sieve depth)

1030T-1045T complete, about 15200 factors found:

[URL]http://www.sendspace.com/file/kmy90g[/URL]

2000T-2010T complete, 5101 factors found:

[URL]http://www.sendspace.com/file/s12sqr[/URL]

It looks like doubling the sieve depth means that you need to take more than twice as much time to find the same number of factors (my earlier 1000T-1005T range had 5307 factors).

740T-900T finished. Taking 1045T-1200T.

[QUOTE=gribozavr;231052]740T-900T finished. Taking 1045T-1200T.[/QUOTE]
You're the first (and only) TPS participant to reach the 1 million and 10 million milestones on the stats page. Congrats!

That also means that all n=480K-485K below 1P have been completely sieved. For reference, we didn't reach the 1P milestone for our first single n sieve until the very end of that effort:

[URL]http://www.mersenneforum.org/showpost.php?p=94124&postcount=49[/URL]

1025T-1030T complete, 4975 factors: [URL]http://www.sendspace.com/file/idogfl[/URL]

Reserving 1200T-1205T.

gribozavr, could you download and verify the factors from the posts in this thread and upload a small k file (k<100,000 for n=481400-484900)?

The LLRnet server dried up, so we don't have any k/n pairs to test.