Gibt es noch Optimierungspotential? (Mega8 an 320x240 GLCD Textmodus)

**MagicWSmoke** · 21.04.2012, 20:00

Zitat von MisterMou

Hmm, von asm hab ich nicht so viel Ahnung, sonst hätte ich gleich alles so geschrieben

So 'ne Routine in ASM ist eigentlich kein Problem, nur beim C Inline Assembler hab' ich so meine Bauchschmerzen.
Aber es lesen zu können reicht hier ja auch schon. Die Rechnerei die der Compiler beim Progmem Zugriff veranstaltet hatte, sagte mir dass es besser ist dort einen Pointer zu verwenden.

Ich habe die neuste AVR-Studio 5 Version instaliert

Und ich die 4er Version.

**MisterMou** · 22.04.2012, 10:39

Zitat von MagicWSmoke

Die innerste und damit wichtigste Loop sieht jetzt so aus, da wüsste ich nicht mehr viel zu optimieren:

Code:

L0673:
    ld      r24,X+
    movw    r30,r18
    add     r30,r24
    adc     r31,r1
    lpm     r30,Z
    out     p18,r25
    out     p15,r30
    out     p18,r1
    out     p18,r25
    swap    r30
    out     p15,r30
    out     p18,r1
    cp      r26,r16
    cpc     r27,r17
    brne    L0673

Was macht diese Zeile?

Code:

 adc     r31,r1

r1 müsste doch 0 sein...
Was der Befehl macht weiß ich(denke ich), aber warum?
r31 verändert sich doch nicht, welche Rolle spielt das Carry?

**sternst** · 22.04.2012, 11:04

Zitat von MisterMou

Was macht diese Zeile?

Sie addiert den Übertrag zu R31, der möglicherweise bei der Zeile vorher entstanden ist.

Code:

    add     r30,r24
    adc     r31,r1

R24 wird zum 16-Bit Wert in R30:31 addiert.

**MagicWSmoke** · 22.04.2012, 11:07

Zitat von MisterMou

Was macht diese Zeile?

Code:

 adc     r31,r1

Sie sorgt dafür, wie Du richtig erkannt hast, dass eine 0 addiert und das Carry übernommen wird.

r31 verändert sich doch nicht, welche Rolle spielt das Carry?

Doch, R31 muss sich ändern, da der Fontblock nicht aligned ist, siehst Du wenn Du Dir den Pointer ausgeben lässt, bei mir startete der bei 0x26.

Wenn Du dem Compiler klarmachen kannst, dass er für den Font auf ein Alignement auf einen 256 Byte Block macht, dann sparst Du Dir den ADC, 1 Takt, mehr aber nicht, da das MOVW gleichviel wie ein MOV braucht.
Und sollte Dir der Code nicht zu hässlich werden und wenn Du genügend Flash über hast, dann kannst Du diese innere Schleife entrollen und einfach 40 mal wiederholen, das spart Dir dann mehr, nämlich 2 Takte pro Durchlauf.

**MisterMou** · 22.04.2012, 13:31

Ahh, genau, der Flash hat ja 16Bit Adressen und das Array liegt bestimmt nicht nur in einem 256Byte Block, damit ist alles klar

Ich sag ja, asm ist nicht meine Welt.

Das mit dem Entrollen hatte ich gestern schon mal probiert, dabei wurde "k" durch Konstanten ersetzt. Das Ergebnis war langsamer o.O

Bild hier

http://www.myvideo.de/watch/8517232/...GLCD_Textmodus

**MagicWSmoke** · 23.04.2012, 00:08

So mag ich das, wenn da passende Grüße kommen

Zitat von MisterMou

Das mit dem Entrollen hatte ich gestern schon mal probiert, dabei wurde "k" durch Konstanten ersetzt. Das Ergebnis war langsamer o.O

Langsamer muss es nicht zwingend sein, mit Konstanten wird's aber nix. Nachdem der Compiler in mancher Beziehung extrem stur ist, hab' ich's nochmal anders versucht.
Wenn's geht, was ich doch hoffe, zumindest die Simulation sah gut aus, so wären das 107Hz@16MHz. Optimierung -Os.

Code:

 /*
     *B0-CP
     *B1-LOAD
     *B2-FRAME
     *
     *C0-D0
     *C1-D1
     *C2-D2
     *C3-D3
     *
     *Mega8 @ 16MHz
     */ 

    #define CP 0
    #define LOAD 1
    #define FRAME 2

    #include <avr/io.h>
    #include <avr/interrupt.h>
    #include <avr/pgmspace.h>

    //8x12 Schrift
    //http://www.mikrocontroller.net/topic/54860
    //Danke Benedikt!

    const unsigned char schrift[12][256] PROGMEM ={
    {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x03,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x03,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xC0,0x03,0x00,0x0C,0x83,0x00,0x03,0x00,0x0C,0x00,0x01,0x06,0x00,0x87,0xC0,0x00,0x00,0x03,0x00,0x0C,0x03,0x0C,0x00,0x00,0xCC,0x00,0xC3,0x00,0x00,0xE0,0xC0,0xC0,0xC0,0xC0,0x00,0x67,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x42,0x55,0xD6,0x81,0x81,0xC0,0x87,0x06,0x00,0x66,0x66,0x00,0x66,0x00,0xCC,0x00,0x81,0x81,0x00,0x81,0x00,0x81,0x00,0x67,0x66,0x00,0x66,0x00,0x66,0x00,0x66,0x00,0xCC,0x00,0x87,0x00,0x06,0x00,0x81,0x87,0x00,0x81,0x00,0xFF,0x00,0x00,0x06,0xFF,0x81,0x00,0x87,0x06,0x00,0x67,0x00,0x00,0x00,0x81,0x87,0x06,0x60,0x81,0x00,0xC0,0x00,0x00,0x00,0x0E,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00},
 {0x00,0xE7,0xE7,0x00,0x01,0x81,0x81,0x00,0x00,0x00,0x00,0xE3,0xC3,0x00,0xF7,0x00,0x08,0x20,0x81,0x66,0xF7,0xE7,0x00,0x81,0x81,0x81,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x03,0x66,0xC6,0x03,0x00,0x07,0x03,0xC0,0x06,0x00,0x00,0x00,0x00,0x00,0x00,0xC7,0x01,0x87,0x87,0xC0,0xCF,0x83,0xEF,0x87,0x87,0x00,0x00,0xC0,0x00,0x06,0x87,0xC7,0x03,0xCF,0xC3,0x8F,0xEF,0xEF,0xC3,0xCC,0x87,0xE1,0x6E,0x0F,0x6C,0x6C,0x83,0xCF,0x83,0xCF,0x87,0xCF,0xCC,0xCC,0x6C,0xCC,0xCC,0xEF,0xC3,0x00,0xC3,0x83,0x00,0x03,0x00,0x0E,0x00,0xC1,0x00,0x83,0x00,0x0E,0x81,0xC0,0x0E,0x87,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xC1,0x81,0x0E,0x37,0x00,0x87,0xCC,0x81,0x87,0xCC,0x06,0xC6,0x00,0x87,0xCC,0x06,0xC6,0x83,0x03,0xCC,0xCC,0x81,0x00,0xE3,0x87,0xCC,0x06,0x87,0x06,0x66,0xCC,0x00,0x00,0x66,0xA3,0x00,0xB1,0x81,0x81,0x81,0x81,0x67,0xCD,0x87,0x87,0x03,0x83,0x00,0x26,0x36,0x03,0x00,0x00,0x29,0xAA,0xBD,0x81,0x81,0x81,0xCC,0x03,0x83,0x66,0x66,0x00,0x66,0x03,0xCC,0x00,0x81,0x81,0x00,0x81,0x00,0x81,0x67,0xCD,0x66,0x00,0x66,0x00,0x66,0x00,0x66,0x00,0x03,0x8F,0xCC,0xCC,0x03,0x0F,0x03,0xCC,0xCC,0x81,0x00,0xFF,0x00,0x81,0x03,0xFF,0x03,0x87,0xCC,0x03,0x67,0xCD,0x00,0x00,0x0F,0x03,0xCC,0x03,0xC0,0x03,0xCF,0x81,0x00,0x00,0x00,0x33,0xF7,0xE7,0x00,0x00,0xC3,0xCC,0x00,0x03,0x87,0x87,0x00,0x00},
 {0x00,0x3C,0xFF,0x44,0x83,0xC3,0xC3,0x00,0x00,0x00,0x00,0xE0,0x66,0x00,0x36,0x81,0x0C,0x60,0xC3,0x66,0xBD,0x36,0x00,0xC3,0xC3,0x81,0x00,0x00,0x00,0x00,0x01,0xEF,0x00,0x87,0x66,0xC6,0xC7,0x00,0x8D,0x03,0x81,0x03,0x00,0x00,0x00,0x00,0x00,0x20,0x6C,0x03,0xCC,0xCC,0xC1,0x0C,0x06,0x6C,0xCC,0xCC,0x00,0x00,0x81,0x00,0x03,0xCC,0x6C,0x87,0x66,0x66,0xC6,0x26,0x66,0x66,0xCC,0x03,0xC0,0x66,0x06,0xEE,0x6C,0xC6,0x66,0xC6,0x66,0xCC,0x4B,0xCC,0xCC,0x6C,0xCC,0xCC,0xEC,0x03,0x08,0xC0,0xC6,0x00,0x81,0x00,0x06,0x00,0xC0,0x00,0xC6,0x00,0x06,0x81,0xC0,0x06,0x81,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x03,0x81,0x03,0xAD,0x00,0xCC,0xCC,0x03,0xCC,0xCC,0x03,0xC6,0x00,0xCC,0xCC,0x03,0xC6,0xC6,0x81,0x00,0xCC,0x00,0x00,0x87,0xCC,0xCC,0x03,0xCC,0x03,0x66,0x00,0xCC,0x00,0x06,0xC6,0x00,0x81,0x03,0x03,0x03,0x03,0xCD,0x00,0xCC,0xCC,0x03,0x44,0x00,0x6E,0x6E,0x03,0x00,0x00,0x94,0x55,0x6B,0x81,0x81,0x00,0x00,0x00,0x44,0x66,0x66,0x00,0x66,0x03,0xCC,0x00,0x81,0x81,0x00,0x81,0x00,0x81,0xCD,0x00,0x66,0x00,0x66,0x00,0x66,0x00,0x66,0x00,0x8D,0xC6,0x00,0x00,0x00,0x03,0x00,0x00,0x00,0x81,0x00,0xFF,0x00,0x81,0x00,0xFF,0x00,0xCC,0x00,0x00,0xCD,0x00,0x00,0x0E,0x06,0x00,0x00,0x00,0x81,0x00,0x00,0x03,0x00,0x03,0x00,0x66,0xBD,0x36,0x03,0x00,0x66,0x00,0x00,0x07,0xC0,0xC0,0x00,0x00},
 {0x00,0x18,0xFF,0xEE,0xC7,0xC3,0xE7,0x00,0x00,0x00,0x00,0xA3,0x66,0x00,0xF7,0xBD,0x0E,0xE0,0xE7,0x66,0xBD,0x03,0x00,0xE7,0xE7,0x81,0x81,0x03,0x00,0x42,0x01,0xEF,0x00,0x87,0x66,0xEF,0x0C,0x4C,0x8D,0x03,0x03,0x81,0x66,0x81,0x00,0x00,0x00,0x60,0xEC,0x0F,0xCC,0xC0,0xC3,0x0C,0x0C,0x6C,0xCC,0xCC,0x83,0x83,0x03,0x00,0x81,0xC0,0x6C,0xCC,0x66,0x6C,0x66,0x06,0x26,0x6C,0xCC,0x03,0xC0,0xC6,0x06,0xEF,0x6E,0x6C,0x66,0x6C,0x66,0xCC,0x03,0xCC,0xCC,0x6C,0xCC,0xCC,0x89,0x03,0x0C,0xC0,0x6C,0x00,0x00,0x00,0x06,0x00,0xC0,0x00,0x06,0x00,0x06,0x00,0x00,0x06,0x81,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x06,0x00,0x00,0x00,0x00,0x00,0x00,0x03,0x81,0x03,0xEC,0x01,0xCC,0x00,0x00,0x00,0x00,0x00,0x83,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x03,0x87,0xCF,0x00,0x8D,0x00,0x00,0x00,0x00,0x00,0x00,0x87,0xCC,0x00,0x06,0xEC,0x00,0x81,0x00,0x00,0x00,0x00,0x00,0x6C,0xCC,0xCC,0x00,0xAB,0x00,0xC6,0xC6,0x00,0x00,0x00,0x42,0xAA,0xD6,0x81,0x81,0x03,0x03,0x03,0xAB,0x66,0x66,0x00,0x66,0x87,0xCC,0x00,0x81,0x81,0x00,0x81,0x00,0x81,0x00,0x03,0x66,0x00,0x66,0x00,0x66,0x00,0x66,0x00,0xC0,0x66,0xCF,0xCF,0xCF,0x03,0x87,0x87,0x87,0x81,0x00,0xFF,0x00,0x81,0x87,0xFF,0x87,0xCC,0x87,0x87,0x00,0x87,0x00,0x06,0xC7,0xCC,0xCC,0xCC,0x00,0xCC,0x00,0x00,0x00,0x03,0x00,0xC3,0xBD,0x03,0x03,0x00,0x66,0x00,0x00,0x03,0x83,0x81,0xCF,0x00},
 {0x00,0x5A,0xBD,0xEF,0xEF,0xFF,0xFF,0x00,0x00,0x00,0x00,0x27,0x66,0x00,0x36,0xE7,0x8F,0xE3,0x81,0x66,0xBD,0xC3,0x00,0x81,0x81,0x81,0xC0,0x06,0x0C,0x66,0x83,0xC7,0x00,0x87,0x42,0xC6,0x0C,0xCC,0x07,0x06,0x06,0xC0,0xC3,0x81,0x00,0x00,0x00,0xC0,0xED,0x03,0xC0,0xC0,0xC6,0x0C,0x0C,0x60,0xCC,0xCC,0x83,0x83,0x06,0xE7,0xC0,0x81,0xED,0xCC,0x66,0x0C,0x66,0x46,0x46,0x0C,0xCC,0x03,0xC0,0xC6,0x06,0xEF,0x6F,0x6C,0x66,0x6C,0x66,0x0C,0x03,0xCC,0xCC,0x6C,0x87,0xCC,0x81,0x03,0x06,0xC0,0x00,0x00,0x00,0x87,0xC7,0x87,0xC7,0x87,0x06,0x67,0xC6,0x87,0xC3,0x66,0x81,0xCF,0x8F,0x87,0xCD,0x67,0xCE,0x87,0xCF,0xCC,0xCC,0x6C,0x6C,0x66,0xCF,0x06,0x81,0x81,0x00,0x83,0x0C,0xCC,0x87,0x87,0x87,0x87,0x8F,0x87,0x87,0x87,0x87,0x87,0x87,0x87,0x87,0x87,0x4C,0xEF,0x8D,0x87,0x87,0x87,0xCC,0xCC,0x66,0xCC,0xCC,0x87,0x06,0x6D,0x00,0xE7,0x87,0x87,0x87,0xCC,0x8F,0x6E,0xE7,0x87,0x03,0xAA,0x00,0x87,0x87,0x03,0x33,0xCC,0x29,0x55,0xBD,0x81,0x81,0x87,0x87,0x87,0x2A,0x6E,0x66,0xEF,0x6E,0xCC,0x87,0x00,0x81,0x81,0x00,0x81,0x00,0x81,0x87,0x87,0x76,0xF7,0x7E,0xFF,0x76,0xFF,0x7E,0x00,0x60,0x66,0x4C,0x4C,0x4C,0x03,0x03,0x03,0x03,0x81,0x00,0xFF,0x00,0x81,0x03,0xFF,0xCC,0x8D,0xCC,0xCC,0x87,0xCC,0x66,0xC7,0x66,0xCC,0xCC,0xCC,0x66,0xCC,0x00,0x00,0x00,0xCF,0x00,0x8F,0xBD,0xC3,0x00,0x00,0x66,0x00,0x00,0x03,0xC0,0x03,0xCF,0x00},
 {0x00,0x18,0xFF,0xEF,0xEF,0x7E,0xFF,0x00,0x00,0x00,0x00,0x8F,0xC3,0x00,0x36,0x7E,0xEF,0xEF,0x81,0x66,0xB7,0x66,0x00,0x81,0x81,0x81,0xEF,0xEF,0x0C,0xFF,0x83,0xC7,0x00,0x03,0x00,0xC6,0x87,0x81,0xAF,0x00,0x06,0xC0,0xFF,0xE7,0x00,0xEF,0x00,0x81,0x6D,0x03,0x81,0x83,0xCC,0x8F,0x8F,0xC0,0x87,0xC7,0x00,0x00,0x0C,0x00,0x60,0x03,0xED,0xCC,0xC7,0x0C,0x66,0xC7,0xC7,0x0C,0xCF,0x03,0xC0,0x87,0x06,0x6D,0xEF,0x6C,0xC7,0x6C,0xC7,0x07,0x03,0xCC,0xCC,0x6D,0x03,0x87,0x03,0x03,0x03,0xC0,0x00,0x00,0x00,0xC0,0x66,0xCC,0xCC,0xCC,0x8F,0xCC,0x67,0x81,0xC0,0xC6,0x81,0x6D,0xCC,0xCC,0x66,0xCC,0xE6,0xCC,0x06,0xCC,0xCC,0x6C,0xC6,0x66,0xC8,0x0C,0x00,0xC0,0x00,0xC6,0x0C,0xCC,0xCC,0xC0,0xC0,0xC0,0xC0,0xCC,0xCC,0xCC,0xCC,0x81,0x81,0x81,0xCC,0xCC,0x0C,0xB1,0xEF,0xCC,0xCC,0xCC,0xCC,0xCC,0x66,0xCC,0xCC,0xCC,0xCF,0x6D,0x6C,0x81,0xC0,0x81,0xCC,0xCC,0xCC,0x6F,0x00,0x00,0x06,0xAB,0xCF,0x03,0x73,0x03,0x66,0x66,0x94,0xAA,0x6B,0x81,0x8F,0xCC,0xCC,0xCC,0x2A,0x60,0x66,0x60,0x60,0x0C,0xCF,0x8F,0xF1,0xFF,0xFF,0xF1,0xFF,0xFF,0xC0,0xCC,0x06,0x06,0x00,0x00,0x06,0x00,0x00,0x6C,0xE7,0x6F,0x0C,0x0C,0x0C,0xCF,0x03,0x03,0x03,0x8F,0xF1,0xFF,0x00,0x00,0x03,0xFF,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,0x66,0x66,0x66,0xCC,0xCC,0xCC,0x66,0xCC,0x00,0x00,0xCF,0x03,0x8F,0x73,0xB7,0x66,0xCF,0x00,0xC3,0x00,0x81,0x87,0x87,0xC7,0xCF,0x00},
 {0x00,0xDB,0x3C,0xEF,0xC7,0x7E,0xE7,0x00,0x00,0x00,0x00,0xCC,0x81,0x00,0x36,0x7E,0x8F,0xE3,0x81,0x00,0xB1,0x66,0x00,0x81,0x81,0x81,0xC0,0x06,0x0C,0x66,0xC7,0x83,0x00,0x03,0x00,0xC6,0xC0,0x03,0xED,0x00,0x06,0xC0,0xC3,0x81,0x00,0x00,0x00,0x03,0x6F,0x03,0x03,0xC0,0xEF,0xC0,0xCC,0x81,0xCC,0x81,0x00,0x00,0x06,0xE7,0xC0,0x03,0xED,0xCF,0x66,0x0C,0x66,0x46,0x46,0xEC,0xCC,0x03,0xCC,0xC6,0x26,0x6C,0xED,0x6C,0x06,0xEC,0xC6,0x81,0x03,0xCC,0xCC,0x6D,0x87,0x03,0x06,0x03,0x81,0xC0,0x00,0x00,0x00,0xC7,0x66,0x0C,0xCC,0xCF,0x06,0xCC,0x66,0x81,0xC0,0x87,0x81,0x6D,0xCC,0xCC,0x66,0xCC,0x67,0x06,0x06,0xCC,0xCC,0x6D,0x83,0x66,0x81,0x06,0x81,0x81,0x00,0x6C,0x0C,0xCC,0xCF,0xC7,0xC7,0xC7,0xC7,0x0C,0xCF,0xCF,0xCF,0x81,0x81,0x81,0xCC,0xCC,0x8F,0xF7,0x8D,0xCC,0xCC,0xCC,0xCC,0xCC,0x66,0xCC,0xCC,0xCD,0x06,0x6D,0xC6,0x81,0xC7,0x81,0xCC,0xCC,0xCC,0xED,0xEF,0xEF,0x0C,0x2B,0xC0,0xE6,0xF6,0x87,0xCC,0x33,0x42,0x55,0xD6,0x81,0x81,0xCC,0xCC,0xCC,0x2A,0x60,0x66,0x60,0x60,0x0C,0x03,0x81,0x00,0x00,0x81,0x81,0x00,0x81,0xC7,0xCC,0x06,0x06,0x00,0x00,0x06,0x00,0x00,0xC7,0x6C,0x66,0x8F,0x8F,0x8F,0x00,0x03,0x03,0x03,0x00,0x81,0xFF,0xFF,0x81,0x03,0x00,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,0x66,0x66,0x66,0xCC,0xCC,0xCC,0x66,0x87,0x00,0x00,0x00,0x03,0x00,0xF6,0xB1,0x66,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xCF,0x00},
 {0x00,0x99,0x7E,0xC7,0x83,0x81,0x81,0x00,0x00,0x00,0x00,0xCC,0xE7,0x00,0x76,0xE7,0x0E,0xE0,0xE7,0x00,0xB1,0xC3,0xEF,0xE7,0x81,0xE7,0x81,0x03,0xEF,0x42,0xC7,0x83,0x00,0x00,0x00,0xEF,0xC0,0x06,0xCC,0x00,0x03,0x81,0x66,0x81,0x00,0x00,0x00,0x06,0x6E,0x03,0x06,0xC0,0xC0,0xC0,0xCC,0x03,0xCC,0x81,0x83,0x83,0x03,0x00,0x81,0x00,0x0C,0xCC,0x66,0x6C,0x66,0x06,0x06,0x6C,0xCC,0x03,0xCC,0xC6,0x66,0x6C,0xEC,0x6C,0x06,0xED,0x66,0xCC,0x03,0xCC,0xCC,0xC6,0xCC,0x03,0x26,0x03,0xC0,0xC0,0x00,0x00,0x00,0xCC,0x66,0x0C,0xCC,0x0C,0x06,0xCC,0x66,0x81,0xC0,0xC6,0x81,0x6D,0xCC,0xCC,0x66,0xCC,0x06,0x81,0x06,0xCC,0xCC,0x6D,0x83,0x66,0x06,0x03,0x81,0x03,0x00,0x6C,0xCC,0xCC,0x0C,0xCC,0xCC,0xCC,0xCC,0x0C,0x0C,0x0C,0x0C,0x81,0x81,0x81,0xCF,0xCF,0x0C,0x8D,0x8D,0xCC,0xCC,0xCC,0xCC,0xCC,0x66,0xCC,0xCC,0xCE,0x06,0x6E,0x83,0x81,0xCC,0x81,0xCC,0xCC,0xCC,0xEC,0x00,0x00,0x0C,0xAA,0xC0,0x3C,0xBD,0x87,0xCC,0x33,0x29,0xAA,0xBD,0x81,0x81,0xCF,0xCF,0xCF,0xAB,0x6E,0x66,0x6E,0xEF,0xCC,0xCF,0x81,0x00,0x00,0x81,0x81,0x00,0x81,0xCC,0xCF,0xF7,0x76,0xFF,0x7E,0x76,0xFF,0x7E,0xC6,0x6C,0x66,0x0C,0x0C,0x0C,0x00,0x03,0x03,0x03,0x00,0x81,0xFF,0xFF,0x81,0x03,0x00,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,0x66,0xC7,0xC7,0xCC,0xCC,0xCC,0x66,0x03,0x00,0x00,0x00,0x00,0x00,0xBD,0xB1,0xC3,0x03,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xCF,0x00},
 {0x00,0x3C,0xFF,0x83,0x01,0x81,0x81,0x00,0x00,0x00,0x00,0xCC,0x81,0x00,0x7E,0xBD,0x0C,0x60,0xC3,0x66,0xB1,0xC0,0xEF,0xC3,0x81,0xC3,0x00,0x00,0x00,0x00,0xEF,0x01,0x00,0x03,0x00,0xC6,0x8F,0xCC,0xCD,0x00,0x81,0x03,0x00,0x00,0x83,0x00,0x83,0x0C,0x6C,0x03,0xCC,0xCC,0xC0,0xCC,0xCC,0x03,0xCC,0x03,0x83,0x83,0x81,0x00,0x03,0x03,0x0C,0xCC,0x66,0x66,0xC6,0x26,0x06,0x66,0xCC,0x03,0xCC,0x66,0x66,0x6C,0x6C,0xC6,0x06,0xC7,0x66,0xCC,0x03,0xCC,0x87,0xC6,0xCC,0x03,0x6C,0x03,0x60,0xC0,0x00,0x00,0x00,0xCC,0x66,0xCC,0xCC,0xCC,0x06,0xC7,0x66,0x81,0xC0,0x66,0x81,0x6D,0xCC,0xCC,0x66,0xCC,0x06,0xCC,0xC6,0xCC,0x87,0xC6,0xC6,0xC3,0x4C,0x03,0x81,0x03,0x00,0xEF,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,0xCC,0x0C,0x0C,0x0C,0x81,0x81,0x81,0xCC,0xCC,0x4C,0x8D,0x8D,0xCC,0xCC,0xCC,0xCC,0xCC,0xC3,0xCC,0xCC,0xCC,0x0C,0xC6,0xC6,0x8D,0xCC,0x81,0xCC,0xCC,0xCC,0x6C,0x00,0x00,0xCC,0x44,0xC0,0x68,0x3B,0x87,0x66,0x66,0x94,0x55,0x6B,0x81,0x81,0xCC,0xCC,0xCC,0x44,0x66,0x66,0x66,0x00,0x87,0x03,0x81,0x00,0x00,0x81,0x81,0x00,0x81,0xCC,0xCC,0x00,0x66,0x00,0x66,0x66,0x00,0x66,0xC7,0x6C,0xC6,0x4C,0x4C,0x4C,0x00,0x03,0x03,0x03,0x00,0x81,0xFF,0xFF,0x81,0x03,0x00,0xCC,0x8F,0xCC,0xCC,0xCC,0xCC,0x66,0x06,0x06,0xCC,0xCC,0xCC,0xC3,0x03,0x00,0x00,0x00,0xCF,0x8F,0x3B,0xB1,0xC0,0x03,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xCF,0x00},
 {0x00,0xE7,0xE7,0x01,0x00,0xE7,0xE7,0x00,0x00,0x00,0x00,0x87,0x81,0x00,0x6E,0x81,0x08,0x20,0x81,0x66,0xB1,0x6C,0xEF,0x81,0x81,0x81,0x00,0x00,0x00,0x00,0xEF,0x01,0x00,0x03,0x00,0xC6,0x03,0xC8,0x67,0x00,0xC0,0x06,0x00,0x00,0x83,0x00,0x83,0x08,0xC7,0xCF,0xCF,0x87,0xE1,0x87,0x87,0x03,0x87,0x07,0x00,0x81,0xC0,0x00,0x06,0x03,0xC7,0xCC,0xCF,0xC3,0x8F,0xEF,0x0F,0xE3,0xCC,0x87,0x87,0x6E,0xEF,0x6C,0x6C,0x83,0x0F,0xC0,0x6E,0x87,0x87,0x87,0x03,0xC6,0xCC,0x87,0xEF,0xC3,0x20,0xC3,0x00,0x00,0x00,0x67,0xCD,0x87,0x67,0x87,0x0F,0xC0,0x6E,0xE7,0xCC,0x6E,0xE7,0x6C,0xCC,0x87,0xC7,0xC7,0x0F,0x87,0x83,0x67,0x03,0xC6,0x6C,0xC0,0xCF,0xC1,0x81,0x0E,0x00,0x00,0x87,0x67,0x87,0x67,0x67,0x67,0x67,0x87,0xC7,0xC7,0xC7,0xE7,0xE7,0xE7,0xCC,0xCC,0xCF,0xFE,0xED,0x87,0x87,0x87,0x67,0x67,0xC0,0x87,0x87,0x87,0xEF,0x8B,0x6C,0x07,0x67,0xE7,0x87,0x67,0xCC,0x6C,0x00,0x00,0x87,0x83,0x00,0xC0,0xF3,0x03,0x33,0xCC,0x42,0xAA,0xD6,0x81,0x81,0xCC,0xCC,0xCC,0x83,0x66,0x66,0x66,0x00,0x03,0x03,0x81,0x00,0x00,0x81,0x81,0x00,0x81,0x67,0xCC,0x00,0x66,0x00,0x66,0x66,0x00,0x66,0x6C,0xC7,0x8F,0xCF,0xCF,0xCF,0x00,0x87,0x87,0x87,0x00,0x81,0xFF,0xFF,0x81,0x87,0x00,0x87,0x0C,0x87,0x87,0x87,0x87,0xB7,0x0F,0x0F,0x87,0x87,0x87,0xC0,0x87,0x00,0x00,0x00,0x00,0x00,0xF3,0xB1,0x6C,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00},
 {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x0C,0x00,0x00,0x00,0x00,0x00,0x00,0xE7,0x00,0xE7,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x03,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x06,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x03,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xE1,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xFF,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xCC,0x00,0x00,0xCC,0x00,0x00,0x00,0x00,0x00,0x06,0xC0,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x81,0x00,0x00,0x00,0x00,0x00,0x00,0x03,0x00,0x00,0x00,0x00,0x00,0x00,0x03,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x81,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0xF1,0x30,0x00,0x00,0x00,0x29,0x55,0xBD,0x81,0x81,0x00,0x00,0x00,0x00,0x66,0x66,0x66,0x00,0x03,0x00,0x81,0x00,0x00,0x81,0x81,0x00,0x81,0x00,0x00,0x00,0x66,0x00,0x66,0x66,0x00,0x66,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x81,0xFF,0xFF,0x00,0x00,0x00,0x00,0x06,0x00,0x00,0x00,0x00,0x06,0x00,0x00,0x00,0x00,0x00,0x81,0x00,0x00,0x00,0x00,0x00,0x00,0x30,0x00,0xE7,0x00,0x03,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00},
 {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x87,0x00,0x00,0x87,0x00,0x00,0x00,0x00,0x00,0x0F,0xE1,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x0F,0x00,0x00,0x00,0x00,0x00,0x00,0x06,0x00,0x00,0x00,0x00,0x00,0x00,0x06,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x0F,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x94,0xAA,0x6B,0x81,0x81,0x00,0x00,0x00,0x00,0x66,0x66,0x66,0x00,0x00,0x00,0x81,0x00,0x00,0x81,0x81,0x00,0x81,0x00,0x00,0x00,0x66,0x00,0x66,0x66,0x00,0x66,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x81,0xFF,0xFF,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x0C,0x00,0x00,0x00,0x00,0x00,0x0F,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x07,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00}
    };

    #define char_put() \
      {asm volatile("ld __tmp_reg__, x+"      "\n\t" \
                    "movw  r30, %A0"          "\n\t" \
                    "add   r30, __tmp_reg__"  "\n\t" \
                    "adc   r31, __zero_reg__" "\n\t" \
                    "lpm   %A1, z"            "\n\t" \
                    : \
                    : "r" (&schrift[j][0]), "r" (temp) \
                    : "r26", "r27", "r30", "r31" \
                   ); \
            PORTB=(1<<CP); \
            PORTC=temp; \
            PORTB=0; \
            PORTB=(1<<CP); \
            asm volatile ("swap %0" : "=r" (temp) : "0" (temp)); \
            PORTC=temp; \
            PORTB=0; \
       }


    int main(void)
    {
        unsigned char text[20][40];
        unsigned char zeichen = 0;
        volatile unsigned char dummy;
        unsigned char temp = 0;
        unsigned char *tptr;
    
        for (unsigned char y = 0; y < 20; y++)
        {
            for (unsigned char x = 0; x < 40; x++)
                   text[y][x] = zeichen++;
        }
            dummy = text[19][39];  // optimierung verhindern
    
        DDRC=0x0F;
        DDRB=0x07;
    
    
        PORTB=(1<<FRAME)|(1<<LOAD);       
    
        while(1)
        {
        tptr = &text[0][0];
            PORTB=(1<<FRAME)|(1<<CP);            
            for(volatile unsigned char lines =0; lines<20;lines++)
            {
                for(volatile unsigned char rows =0; rows<12;rows++)
                {
                unsigned char j = rows;
                    asm volatile("movw r26, %A0" : : "r" (tptr) : "r26", "r27");
                    char_put() char_put() char_put() char_put();
                    char_put() char_put() char_put() char_put();
                    char_put() char_put() char_put() char_put();
                    char_put() char_put() char_put() char_put();
                    char_put() char_put() char_put() char_put();
                    char_put() char_put() char_put() char_put();
                    char_put() char_put() char_put() char_put();
                    char_put() char_put() char_put() char_put();
                    char_put() char_put() char_put() char_put();
                    char_put() char_put() char_put() char_put();            
                PORTB=(1<<LOAD);    
                }
            tptr += 40;
            }
            PORTB=(1<<LOAD)|(1<<FRAME);
        }
    }

Nachtrag:
So werden's 114Hz:

Code:

      const unsigned char schrift[12][256] PROGMEM __attribute__((aligned(256))) ={
      //  ...
    #define char_put() \
      {asm volatile("ld __tmp_reg__, x+"      "\n\t" \
                    "movw  r30, %A0"          "\n\t" \
                    "add   r30, __tmp_reg__"  "\n\t" \
                    "lpm   %A1, z"            "\n\t" \
                    : \
                    : "r" (&schrift[j][0]), "r" (temp) \
                    : "r26", "r27", "r30", "r31" \
                   ); \
            PORTB=(1<<CP); \
            PORTC=temp; \
            PORTB=0; \
            PORTB=(1<<CP); \
            asm volatile ("swap %0" : "=r" (temp) : "0" (temp)); \
            PORTC=temp; \
            PORTB=0; \
       }

**MisterMou** · 23.04.2012, 16:25

Oh man...
Jetzt ist es vorbei, bis gestern war der Code kein Problem, jetzt muss ich erst einmal rausbekommen was was macht

Woher hast Du das Wissen? Gibt es da ein Zauberbuch?

Das Display macht keine Probleme, es läuft auch noch bei 132Hz, jetzt sind es 114,6Hz mit -O2 bzw. -O3.

**MagicWSmoke** · 23.04.2012, 16:44

Zitat von MisterMou

Woher hast Du das Wissen? Gibt es da ein Zauberbuch?

286/386er in Assembler geschrieben, auch dazu gibt's natürlich Bücher, und dann irgendwann mal auf AVR. Da kennt man das Meiste schon und braucht eigentlich nur noch ein Instruction Set Datenblatt.
Mehr Kopfzerbrechen hatte mir der C-Inlineassembler gemacht, der recht kryptisch ist und auch durchaus kreativ Code generiert hatte. Manchmal unfunktionell, aber immer wieder interessant. Für den Assembler gibt's Quellen im Netz, so auch hier im Forum, und manches bringt man durch Trial & Error raus.

Das Display macht keine Probleme, es läuft auch noch bei 132Hz, jetzt sind es 114,6Hz mit -O2 bzw. -O3.

So wie ich's verstehe läuft die aktuelle Hardware mit dem neuen Code, ist das richtig ? Konnte nur "trocken" probieren, war also nicht sicher.

**MisterMou** · 23.04.2012, 16:58

Du bist also ein "alter" Hase

Ich fang ja erst an

Normalerweise wüsste ich nicht einmal was was ein AVR ist

2.Semester Elektrotechnik...

Genau, läuft alles wunderbar, hatte bloß mal zum Test 18,4MHz drauf gegeben

Ich schlage schon Alarm, wenns nicht läuft^^

**MagicWSmoke** · 23.04.2012, 19:32

Ja, ich mach' das schon 'ne geraume Zeit

Ist in Ordnung, schaut doch auch ganz gut aus.
Hilft außerdem immer mit den Stoff weiter zu sein als der Rest.

Ich denke, dass es noch schneller geht, aber dann müsste man die Schleifen komplett in Assembler machen, da ließe sich noch was sparen.
Wie's jetzt ist, find' ich 'ne gute Mischung, es ist noch weitestgehend C und da optimiert, wo der Compiler selbst nicht wollte - zum Teil kam da schon wildes Zeug raus.