WP 34S vs. DM42 decimal128 differences?

Post here to share useful tips and tricks, to ask questions about using your DM42 or to report software-related problems
Post Reply
User avatar
rkf
Posts: 6
Joined: Mon Mar 05, 2018 9:14 am
Location: Stade, Germany

WP 34S vs. DM42 decimal128 differences?

Post by rkf » Tue Mar 20, 2018 9:26 am

(Crosspost from MoHPC HP Forums):

Today I stumbled about a footnote in Walter's WP 34S Owner's manual, where at Page 319 the result of the Calculator Forensics Test is mentioned. To my big surprise, the DP difference between the test result, and 9, is

for WP 34S: -6.2465E-29
for DM42: -6.2466E-29

But why the differerence of 1 ULP?? For the other example (1.0000001^(2^27)), there isn't any difference between the two models, BTW.
Last edited by rkf on Fri Mar 23, 2018 8:43 am, edited 1 time in total.
DM42 S/N 00761

User avatar
Walter
Posts: 726
Joined: Tue May 02, 2017 9:13 am
Location: Close to FRA, Germany

Re: WP 34S vs. DM42 decimal128 differences?

Post by Walter » Tue Mar 20, 2018 10:21 am

If I've understood said test correctly then its results point to different processors and firmware used in different physical calculators. So a slight difference in the results between WP 34S and DM42 wouldn't bother me since both parameters are definitively different.

EDIT: The WP 34S Emulator returns the same as the calculator does: so the WP 34S result is a matter of SW/FW only.
DM42 SN: 00041 --- Follower of Platon.

HP-35, HP-45, ..., HP-50, WP 34S, WP 31S, DM16L

whuyse
Posts: 19
Joined: Thu Dec 21, 2017 12:23 pm

Re: WP 34S vs. DM42 decimal128 differences?

Post by whuyse » Wed Mar 21, 2018 7:55 am

comparison of the intermediate results between Free42 and WP-34S:

Code: Select all

 DEG	Free42						WP-34S
 9
 SIN    1.564344650402308690101053194671668-01  1.564344650402308690101053194671669-01
 COS    9.999962727428850241175162050113502-01  =
 TAN    1.745499985548866079139414092834847-02  =
 ATAN   9.999962727428850241175162050113503-01  =
 ACOS   1.564344650402308690101053194660901-01  =
 ASIN   8.999999999999999999999999999937534 00  8.999999999999999999999999999937535 00
Where the two are different, the result of the WP-34S is correct to all digits.

Cheers, Werner
42S #3249S01123
DM42 #00345

User avatar
Walter
Posts: 726
Joined: Tue May 02, 2017 9:13 am
Location: Close to FRA, Germany

Re: WP 34S vs. DM42 decimal128 differences?

Post by Walter » Wed Mar 21, 2018 9:47 am

whuyse wrote:
Wed Mar 21, 2018 7:55 am
Where the two are different, the result of the WP-34S is correct to all digits.
Kudos to Pauli! :D
DM42 SN: 00041 --- Follower of Platon.

HP-35, HP-45, ..., HP-50, WP 34S, WP 31S, DM16L

toml_12953
Posts: 422
Joined: Wed May 03, 2017 5:46 pm
Location: Malone, NY USA

Re: WP 34S vs. DM42 decimal128 differences?

Post by toml_12953 » Wed Mar 21, 2018 4:00 pm

whuyse wrote:
Wed Mar 21, 2018 7:55 am
comparison of the intermediate results between Free42 and WP-34S:

Code: Select all

 DEG	Free42						WP-34S
 9
 SIN    1.564344650402308690101053194671668-01  1.564344650402308690101053194671669-01
 COS    9.999962727428850241175162050113502-01  =
 TAN    1.745499985548866079139414092834847-02  =
 ATAN   9.999962727428850241175162050113503-01  =
 ACOS   1.564344650402308690101053194660901-01  =
 ASIN   8.999999999999999999999999999937534 00  8.999999999999999999999999999937535 00
Where the two are different, the result of the WP-34S is correct to all digits.

Cheers, Werner
By correct, do you mean the last digit itself is correct or that the answer is correct when rounded to that many places?

Example:

2/3 to 6 places:
.666666 All digits correctly truncated.
.666667 Correctly rounded to 6 places.

Both are "correct" for the way they're calculated.

In SIN and ASIN above, it looks like the 34S rounds while Free42 truncates. If we could see one more digit of the true answer, we'd know.
Tom L

आपका दिन शुभ हो
DM42 SN: 00025 (Beta)
SN: 00221 (Shipping)

whuyse
Posts: 19
Joined: Thu Dec 21, 2017 12:23 pm

Re: WP 34S vs. DM42 decimal128 differences?

Post by whuyse » Wed Mar 21, 2018 4:43 pm

I mean 'correctly rounded'.
9 SINDEG ends in ..668923 so the correctly rounded result is ..669
ASINDEG(1.564344650402308690101053194660901-01) ends in ..535301, so ..535 is the correct answer, and it's not a matter of rounding vs. truncating.

Free42 is occasionally an ULP wrong because it doesn't have the 39-digit intermediate format the WP-34S has. Well, that's my guess.

Cheers, Werner
42S #3249S01123
DM42 #00345

Post Reply