Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(38)

Side by Side Diff: conformance/more/performance/CPUvsGPU.html

Issue 8342021: Add webgl conformance tests r15841. (Closed) Base URL: svn://chrome-svn/chrome/trunk/deps/third_party/webgl/sdk/tests/
Patch Set: Created 9 years, 2 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View unified diff | Download patch | Annotate | Revision Log
« no previous file with comments | « conformance/more/index.html ('k') | conformance/more/performance/bandwidth.html » ('j') | no next file with comments »
Toggle Intra-line Diffs ('i') | Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Property Changes:
Added: svn:eol-style
+ LF
OLDNEW
(Empty)
1 <!DOCTYPE html>
2 <html><head>
3 <meta charset="utf-8">
4 <!--
5 Tests for the OpenGL ES 2.0 HTML Canvas context
6
7 Copyright (C) 2011 Ilmari Heikkinen <ilmari.heikkinen@gmail.com>
8
9 Permission is hereby granted, free of charge, to any person
10 obtaining a copy of this software and associated documentation
11 files (the "Software"), to deal in the Software without
12 restriction, including without limitation the rights to use,
13 copy, modify, merge, publish, distribute, sublicense, and/or sell
14 copies of the Software, and to permit persons to whom the
15 Software is furnished to do so, subject to the following
16 conditions:
17
18 The above copyright notice and this permission notice shall be
19 included in all copies or substantial portions of the Software.
20
21 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
22 EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
23 OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
24 NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
25 HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
26 WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
27 FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
28 OTHER DEALINGS IN THE SOFTWARE.
29
30 -->
31
32 <link rel="stylesheet" type="text/css" href="../unit.css" />
33 <script type="application/x-javascript" src="../unit.js"></script>
34 <script type="application/x-javascript" src="../util.js"></script>
35 <script type="application/x-javascript">
36
37 Tests.autorun = false;
38 Tests.message = "This might take a few seconds to run"
39
40 Tests.startUnit = function() {
41 var gl = document.getElementById('gl').getContext(GL_CONTEXT_ID);
42 var ctx = document.getElementById('2d').getContext('2d');
43 return [gl, ctx];
44 }
45
46 var kernel = [0.006, 0.061, 0.242, 0.383, 0.242, 0.061, 0.006];
47
48 Tests.testGPU = function(gl, ctx) {
49 Tests.gpuGaussianBlur(gl);
50 }
51
52 Tests.testCPU = function(gl, ctx) {
53 Tests.cpuGaussianBlur(ctx);
54 }
55
56 function hblur(ctx,idata) {
57 var d = idata.data;
58 var res = ctx.createImageData(256,256);
59 var rd = res.data;
60 var sumR=0.0,sumG=0.0,sumB=0.0,sumA=0.0, kv=0.0;
61 var col_offset = 0, row_offset = 0, k4 = 0;
62 for (var y=0; y<idata.height; ++y) {
63 col_offset = y * idata.width * 4;
64 for (var x=3; x<idata.width-3; ++x) {
65 row_offset = col_offset+x*4;
66 sumR=sumG=sumB=sumA=0.0;
67 for (var k=-3; k<4; ++k) {
68 k4 = k * 4;
69 kv = kernel[k+3];
70 sumR += d[row_offset+k4+0] * kv;
71 sumG += d[row_offset+k4+1] * kv;
72 sumB += d[row_offset+k4+2] * kv;
73 sumA += d[row_offset+k4+3] * kv;
74 }
75 rd[row_offset+0] = Math.floor(sumR);
76 rd[row_offset+1] = Math.floor(sumG);
77 rd[row_offset+2] = Math.floor(sumB);
78 rd[row_offset+3] = Math.floor(sumA);
79 }
80 }
81 var xr = 3;
82 for (var y=0; y<idata.height; ++y) {
83 col_offset = y * idata.width * 4;
84 for (var x=0; x<xr; ++x) {
85 row_offset = col_offset+x*4;
86 sumR=sumG=sumB=sumA=0.0;
87 for (var k=-3; k<4; ++k) {
88 if (k+x < 0)
89 k4 = 0;
90 else
91 k4 = k * 4;
92 kv = kernel[k+3];
93 sumR += d[row_offset+k4+0] * kv;
94 sumG += d[row_offset+k4+1] * kv;
95 sumB += d[row_offset+k4+2] * kv;
96 sumA += d[row_offset+k4+3] * kv;
97 }
98 rd[row_offset+0] = Math.floor(sumR);
99 rd[row_offset+1] = Math.floor(sumG);
100 rd[row_offset+2] = Math.floor(sumB);
101 rd[row_offset+3] = Math.floor(sumA);
102 }
103 }
104 var xr = idata.width-3;
105 for (var y=0; y<idata.height; ++y) {
106 col_offset = y * idata.width * 4;
107 for (var x=xr; x<idata.width; ++x) {
108 row_offset = col_offset+x*4;
109 sumR=sumG=sumB=sumA=0.0;
110 for (var k=-3; k<4; ++k) {
111 if (k+x >= idata.width)
112 k4 = (idata.width-x-1)*4;
113 else
114 k4 = k * 4;
115 kv = kernel[k+3];
116 sumR += d[row_offset+k4+0] * kv;
117 sumG += d[row_offset+k4+1] * kv;
118 sumB += d[row_offset+k4+2] * kv;
119 sumA += d[row_offset+k4+3] * kv;
120 }
121 rd[row_offset+0] = Math.floor(sumR);
122 rd[row_offset+1] = Math.floor(sumG);
123 rd[row_offset+2] = Math.floor(sumB);
124 rd[row_offset+3] = Math.floor(sumA);
125 }
126 }
127 return res;
128 }
129
130 function vblur(ctx,idata) {
131 var d = idata.data;
132 var res = ctx.createImageData(256,256);
133 var rd = res.data;
134 var sumR=0.0,sumG=0.0,sumB=0.0,sumA=0.0, kv=0.0;
135 var col_offset = 0, row_offset = 0, kfac = idata.width*4;
136 for (var y=3; y<idata.height-3; ++y) {
137 col_offset = y * idata.width * 4;
138 for (var x=0; x<idata.width; ++x) {
139 row_offset = col_offset+x*4;
140 sumR=sumG=sumB=sumA=0.0;
141 for (var k=-3; k<4; ++k) {
142 k4 = k * kfac;
143 kv = kernel[k+3];
144 sumR += d[row_offset+k4+0] * kv;
145 sumG += d[row_offset+k4+1] * kv;
146 sumB += d[row_offset+k4+2] * kv;
147 sumA += d[row_offset+k4+3] * kv;
148 }
149 rd[row_offset+0] = Math.floor(sumR);
150 rd[row_offset+1] = Math.floor(sumG);
151 rd[row_offset+2] = Math.floor(sumB);
152 rd[row_offset+3] = Math.floor(sumA);
153 }
154 }
155 var yr = 3;
156 for (var y=0; y<yr; ++y) {
157 col_offset = y * idata.width * 4;
158 for (var x=0; x<idata.width; ++x) {
159 row_offset = col_offset+x*4;
160 sumR=sumG=sumB=sumA=0.0;
161 for (var k=-3; k<4; ++k) {
162 if (k+y < 0)
163 k4 = 0;
164 else
165 k4 = k * kfac;
166 kv = kernel[k+3];
167 sumR += d[row_offset+k4+0] * kv;
168 sumG += d[row_offset+k4+1] * kv;
169 sumB += d[row_offset+k4+2] * kv;
170 sumA += d[row_offset+k4+3] * kv;
171 }
172 rd[row_offset+0] = Math.floor(sumR);
173 rd[row_offset+1] = Math.floor(sumG);
174 rd[row_offset+2] = Math.floor(sumB);
175 rd[row_offset+3] = Math.floor(sumA);
176 }
177 }
178 var yr = idata.height-3;
179 for (var y=yr; y<idata.height; ++y) {
180 col_offset = y * idata.width * 4;
181 for (var x=0; x<idata.width; ++x) {
182 row_offset = col_offset+x*4;
183 sumR=sumG=sumB=sumA=0.0;
184 for (var k=-3; k<4; ++k) {
185 if (k+y >= idata.height)
186 k4 = (idata.height-y-1)*kfac;
187 else
188 k4 = k * kfac;
189 kv = kernel[k+3];
190 sumR += d[row_offset+k4+0] * kv;
191 sumG += d[row_offset+k4+1] * kv;
192 sumB += d[row_offset+k4+2] * kv;
193 sumA += d[row_offset+k4+3] * kv;
194 }
195 rd[row_offset+0] = Math.floor(sumR);
196 rd[row_offset+1] = Math.floor(sumG);
197 rd[row_offset+2] = Math.floor(sumB);
198 rd[row_offset+3] = Math.floor(sumA);
199 }
200 }
201 return res;
202 }
203
204 Tests.cpuGaussianBlur = function(ctx) {
205 var s = document.getElementById('cpustat');
206 var t0 = new Date().getTime();
207 ctx.drawImage(document.getElementById('logo'),0,0);
208 var idata = ctx.getImageData(0,0,256,256);
209 for (var i=0; i<1; i++){
210 idata = hblur(ctx,idata);
211 idata = vblur(ctx,idata);
212 }
213 ctx.putImageData(idata, 0, 0);
214 var t1 = new Date().getTime();
215 s.textContent = 'Done! Time: '+(t1-t0)+'ms';
216 }
217
218 Tests.gpuGaussianBlur = function(gl) {
219 var s = document.getElementById('gpustat');
220 var t0 = new Date().getTime();
221
222 var fbo1 = new FBO(gl, 256, 256);
223 var fbo2 = new FBO(gl, 256, 256);
224 var hblur = new Filter(gl, 'identity-vert', 'hblur-frag');
225 var vblur = new Filter(gl, 'identity-vert', 'vblur-frag');
226 var identity = new Filter(gl, 'identity-vert', 'identity-frag');
227 var identityFlip = new Filter(gl, 'identity-flip-vert', 'identity-frag');
228
229 gl.viewport(0,0,256,256);
230 gl.clearColor(0,0,1,1);
231 gl.clear(gl.COLOR_BUFFER_BIT | gl.DEPTH_BUFFER_BIT);
232 gl.disable(gl.DEPTH_TEST);
233 gl.activeTexture(gl.TEXTURE0);
234
235 fbo1.use();
236 var tex = loadTexture(gl, document.getElementById('logo'));
237 gl.bindTexture(gl.TEXTURE_2D, tex);
238 identityFlip.apply(); // draw image
239
240 // gaussian blur
241 for (var i=0; i<1000; i++) {
242 fbo2.use();
243 gl.bindTexture(gl.TEXTURE_2D, fbo1.texture);
244 hblur.apply(function(f){
245 f.uniform1f('width', 256.0);
246 f.uniform1i('Texture', 0);
247 });
248 fbo1.use();
249 gl.bindTexture(gl.TEXTURE_2D, fbo2.texture);
250 vblur.apply(function(f){
251 f.uniform1f('height', 256.0);
252 f.uniform1i('Texture', 0);
253 });
254 }
255
256 gl.bindFramebuffer(gl.FRAMEBUFFER, null);
257 gl.bindTexture(gl.TEXTURE_2D, fbo1.texture);
258 identity.apply(); // draw blurred image on screen
259
260 fbo1.destroy();
261 fbo2.destroy();
262 hblur.destroy();
263 vblur.destroy();
264 identity.destroy();
265 identityFlip.destroy();
266 gl.deleteTexture(tex);
267 checkError(gl, "end");
268 var t1 = new Date().getTime();
269 s.textContent = 'Done! Time: '+(t1-t0)+'ms';
270 }
271
272
273 </script>
274 <script id="identity-vert" type="x-shader/x-vertex">
275
276 attribute vec3 Vertex;
277 attribute vec2 Tex;
278
279 varying vec4 texCoord0;
280 void main()
281 {
282 texCoord0 = vec4(Tex, 0.0, 0.0);
283 gl_Position = vec4(Vertex, 1.0);
284 }
285 </script>
286 <script id="identity-flip-vert" type="x-shader/x-vertex">
287
288 attribute vec3 Vertex;
289 attribute vec2 Tex;
290
291 varying vec4 texCoord0;
292 void main()
293 {
294 texCoord0 = vec4(Tex.s, 1.0-Tex.t, 0.0, 0.0);
295 gl_Position = vec4(Vertex, 1.0);
296 }
297 </script>
298 <script id="identity-frag" type="x-shader/x-fragment">
299
300 precision mediump float;
301
302 uniform sampler2D Texture;
303
304 varying vec4 texCoord0;
305 void main()
306 {
307 gl_FragColor = texture2D(Texture, texCoord0.st);
308 }
309 </script>
310 <script id="hblur-frag" type="x-shader/x-fragment">
311
312 precision mediump float;
313
314 uniform sampler2D Texture;
315 uniform float width;
316
317 varying vec4 texCoord0;
318 void main()
319 {
320 float kernel[7] = float[7](0.006, 0.061, 0.242, 0.383, 0.242, 0.061, 0.006);
321 int i;
322 float step = 1.0 / width;
323 vec4 sum = vec4(0.0);
324 for (i=-3; i<=3; i++) {
325 vec4 tmp = texture2D(Texture, texCoord0.st + vec2(float(i)*step, 0.0));
326 sum = (tmp * kernel[i+3]) + sum;
327 }
328 gl_FragColor = sum;
329 }
330 </script>
331 <script id="vblur-frag" type="x-shader/x-fragment">
332
333 precision mediump float;
334
335 uniform sampler2D Texture;
336 uniform float height;
337
338 varying vec4 texCoord0;
339 void main()
340 {
341 float kernel[7] = float[7](0.006, 0.061, 0.242, 0.383, 0.242, 0.061, 0.006);
342 int i;
343 float step = 1.0 / height;
344 vec4 sum = vec4(0.0);
345 for (i=-3; i<=3; i++) {
346 vec4 tmp = texture2D(Texture, texCoord0.st + vec2(0.0, float(i)*step));
347 sum = (tmp * kernel[i+3]) + sum;
348 }
349 gl_FragColor = sum;
350 }
351 </script>
352 </head><body>
353 <img id="logo" src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAQAAAAEACAYAAA BccqhmAAAgAElEQVR4nO19bWiUZ9b/+BJfsP8oJfXlyUMDNkSSfnvGhgh2Qak6IPRZJ9QPtTIlYMQ2RQ vWHagSxcgiCYTtulsCCinb1mWfriy20pSnLkL1Q2Fqv5RktThhaxseWFGXFQbZdc//w8y55lznPtfLfc 99J6OZCw73PTOZSYz5/c7vvFznSkFjNVZjzduVmusfoLEaq7HmbjUIoLEaax6vBgE0VmPN49UggMZqrH m8GgTQWI01j1eDABqrsebxahBAYzXWPF4NAmisxprHq0EAjdVY83g1CKCxGmserwYBPKarWJyGs2c/hT ff/BjefPNjWL/tV7A7ewRWtQwoS6X61dVk9OtXtQzA0x1H1Wf9fP/78OabH8PZs5/C5OQklEqluf5nN1 bMq0EAdbxmZmZg4otvFcApsOfKKEm8+ebHMPHFt3Dv3r25/lU1VsTVIIA6Wffu3VNgf7rjqB/Q1x6E1N qD8HTHUVjccxoW95yGRb2/haaBP8DSX3wKy0/8GZaNXoNlo9fg/330F/h/H/0F/nPi/+A/J/5PPb/8xJ 9h+Yk/w7MHPoTU6x/Bot7fqs/Cz0+tPehFDA1SePxWgwDmYJVKJbh37x68f+Z/4Of733eDvQLy1I4RBW 4ELwKZAp3fI/ip/efE/4nP0/fwz1t+4s/QNPAHSGV+A2vSv4RU59vO0OLn+9+Hs2c/bRBCna4GAcziKh QK8OabH5fBY/PqnW8rT44empoJ/Cbgo9enZgM+vV9+4s8aCXDDnyn1+kewuOd0magMhLAm/Ut4/8z/wM QX3zbyCXWyGgSQ4CqVSjDxxbd2L7/2YFluv/4RLP3Fp5otP/FndfUlARvoTSTAVYJJAdiM/3xLf/EppD K/MYYQq1oG4M03P4bJycm5/m+a16tBAAmsyclJFcu7PDxKerz6gN9EAD6gt6kAE/BRCUig51fJ8N+n8g rsd/J0x1F4882PoVAozPV/3bxbDQKIaWFML8r7tQch9dKgBnpu3PvbSEACHwI1KgFI8t8nDLCBn/4buC l1wH5Xa9K/hLNnP22ECLO0GgRQ4zJK/HWvKU+fev0jSL3+UTmBVrly8IclAQTezwb/FAn8tBpgI4GfDf 7JKwdg8/4mEsB/f2rHSIAMGiHC7KwGAURcE198C+u3/Sro7TvfhtSOkSrw+8ecBOAiAZcKiCL/fZKBth yAj/w3gZ/+O+nvAEuQUnmxkThMZjUIIMRCmR+I7ZtfhtQLeVjU+9sy8Cugj0ICUVRAVO9PCeC/3/tfq+ fH5+KQ/5QEpN9D6vWPRFXQCA/iXw0C8Fj37t2Ds2c/1YHf/HJZ5hPgc6PA1+5DqgAXAUQFf1wKIKz8t4 FfIwEkUdZv8HTHUTh79tO5/rN4IlaDACyrVCoZgb+457SS+hz06p6qgIphx12YZKCNBGoBv5QIdCmBpB UAVUmUNBtEkMxqEIBhTXzxrezxXxosy1OM8wUS4GRgUgBREoIUfLVIf5+KQNyxv68CSPWPaSrARQRr0r +EiS++nes/mcdyNQiArUKhoCf3iNRH4HuBn3h/iQS4Enj2wIcaOPAxBf+zBz5UII3D+9uIgHv+/37vfx OX/ybvn+of08hU/Y4ZEazf9qtG1SDkahBAZRWL0/Dmmx/r5bxKco96fW42IuB/tFQB2MIAmwLg/fxJma v0F9X7+8T+oveniVUaZgnJwjff/LiRKPRcDQIAgPfP/E8Q+J1vQ+qlQRXrSySwqPe3qsEnTBiA3t9XCV BAITiTBD8lAa4EsBLgIoGfDf4J/vu9/4WhY3kYGjoNhw8fhsOHD5cfH8ur++G+LXD48GH4+f73YXf2CD x74EPYsOfXAe8vgV/7Pb+Q14hgVctAIz/gseY1ARQKBb1zr/nl8h9Rxesv7jntJIHI4YCjJGiqBiwbvQ ZLf/Fpzdn/KEqA5yBsCuDZAx/CwBvvwNCxPHx+bj/cOdUNcHUnwNWdUPrqRPUKQ+p5uLoTYCwNMJaG0v EuuJzfCpfzW2HoWB5yuRz8fP/7VrJthAXh17wkgHv37gXlPsb5L+TLgEfpXzFFAhFCAVMugHr/Zw98qI Gf5wCoEkCAhd38E9WWn/iz5vlN3p9XKziBPXvgQ/j5/vdh6Fgefn98M5SOd8Gj4Y0AV3fCwzO9ANAPAE OaIVkgOZSOd8GdU91wOb8VBt54B3K5XFkxVJSYImpCBNhV2FjBNe8IYHJyUk/yrXsNUhv2KvAr4CMREC VQiwoQG4QsSoAnzBD0UjmQqoEkyABVgCv5Fzbxl+ofg5/vfx8OHz4Ml/NboXS8SyMFuLoTOCGUrb9KDI QUhvu2VAlhx0j5/5JtR26oAX3VDQF8/0keHv3tWqLf4+zZT8V+fQ38SZOApRqAj21KgKoB9MxJhwFU9n MlwAkgdN2flPyePfAh5HI5pQ4enOyG0vGuclggkkE/ub6oEcKtwXYY7tsCA2+8UyZ41jvQSBKWV10QQO mrE3D3wGr4/pN8Ip9fLE7Dz/e/r4Mfvb5AACoEIAQQIIE4cgFCKOBTFqRqIKlyoLQVOGr231n3FxJ+qf 4xyOVymjJAMqiGC9xerNrVnQATbUpRXM5vhZFMlwr7fr7/fZiZmUnk7+1xWnNPAIUxmH5rDcy8uxI+P7 c/9o8PNPSg10czkAAlgjhJAAng2QMfaqFA2A5BBFsSyUCa/UfQuzx/FPAr5eMo9W3Y82sY7tuiEYEWKs CLECAASFWv36wGmGiDBye74dHwRpga6IDhvi2w8fneeT/teE4J4NHfrsH9Q+th5t2VcGuwHaYGOmL77F KpBO+f+R9Z8nPwm8IAlhD0CQOcvQG0kYX1BZiUAFcBs6EAuPRPatMPVwBSqU/9vl8ahJFMV4AIVJjwzW pAAvj335vINVW1b1bDow/K4cWDk90wNdABn5/bP29nFs4ZAdy7dw9uDbbDzLsrNRK4d+8elEqlmlh5Zm ZGl/zYzUfBLykAiQgQ+J4kQMGfyvzGGQpoNW+HEpA6BJPMAdCuwLCJP2/pT72/UONX2X1SkUmn03C+v1 VUBI+GNxIiSAUI4N9/byrbDys0VfBoeKPKQ80nRTBnBPD9J3kN/PcPrYfpt9bAP4sXNBII+59RLE4HW3 krY7iMFiIXEOgNiLEsaGsOspFAEpn/JON+XvmQvH+qf0wEv7LFnUY18Gh4Y1kVTLSB5v0pAVRIAIkASQ CJAP8Gn/Q1JwTwz+IFBXxKAjPvroTr43vh3r17ysIQQaFQCO7T37C3DG6TAvDIBRgrAj5VAYMK4CTgkx OgwKf3SfQC0NJfIht+pL5/SfoT+Y//D6mXBiG1YS+sbNsGr6TXWkkAFYEGfAQ/IYF//7ACHn3QpRFB6a sTkR3R47JmnQBKpRLcOdWteX5KBNNvrYGZmRmYmZkJEIFtTXzxrcrwLlu0uQx4BD0FPxJCSBKQwoA4Eo ISEfC9AjwPQPsCku4KxERg3NKfJ/8C3p/8/hT4N+wNXDkJIPAD9x90VYlAAD98s1pdkQgenOyGh2d64d Hfrj2ximBWCaBUKsHt7y4EwI+5AHzupyvvQbE47U0CWrIP430O/rUH7eB3JANtCUFbOBAqFKjcNw38AX 42+KeASU1Czx74MDHwm0iA70/gPQrenl/a6ss3+nDpT0mgYi0tzwRyAhIRPBreCI8+6NKAbyIBGhY8ON mt1MCTRgSzRgD4y7tzqlv0/JwQisVpRQJIBFyKaZn+5pd1z7/2YLXLD8GPuYAoKgDlp1QWrKEqsGHPr+ Fng3+C0dFRuHTpEtz+7gL8s1g1+Mfv4NKlS1oeAEGGQJyNVmDegRhL1p9YAPy2HIBAApQATODn+QEj+I nRsODhmV5vRfq4rFklAOr9TSSA999/ktdIgP/isZ9fBD8aBz8C30UEEfIB1qRgJQ+A1/868DGMjo7CT1 fe04DOgX/7uwvws8E/Qap/LNAhiECjZcC4B4RI3p83I4Xx/qL8pyEALZGS3ZaaEnghrwE/k8nA5fxWuH 9ofQD4qipwdafK9gdeqwBdAj8qAZ4kfPS3azAzM/NE5AZmhQC49zd5ff64cO0KTE5OaiSAv3QE/7JFm2 XwcxXAQwJXOCCQgLZJKGSb8IY9v4bx8XHl4TnguY2Pj8OzBz4MbB3meQEKVAnAcXj+WON+adiHYasvJY Hd2SOQzWZhuG8LDPdtCXQImgwlPLYS8z0Ej4Y3lkFuAj8hARhLKxK5/d2FJ0INzBoBFIvTCuAS2KXH33 +Sh0KhAJOTkyoUUDX+5pchlUqbwW8KBXySgo7eAFdCkKqBN9/8GC5dumQEOvzjdwCFMc0OHz6sAcI0P4 A3AUmgdxEBfV0iEawE8AakqOCn/5ZU/xhs2PNrGHjjHfj98c1wOb8VpgY61B4AbPChgKav2bL/pnu4ur P8O6/sJ3h4pldtQXYSwESbIoFHwxsBCmOPPQkkTgDo/b//JG/M/Juut/etgsK1K1AoFFQ4EAr8lASQAK gCsPUH2MIAQ6swJYGf738/4O0R7P8sXgiAvvTVCSh9dQJyuZyYJJSUAJfqPsD3eZ5/JhKBa7OPS/prxn ZGLur9LezOHoGhY3m4Ndge7PJT8wP0nYCcIHzzABoR4LWyf8AE/kcfdGkkgMlBXrJ+nNasEQDN9Nvif/ 6YhgJl2Z/WwY9Zf5r9d5GABH6XEvBUAQh8KvElsFPQl746AQ/P9MLu7BExWciVAEp/Cai+ZMDBzUuJ0m diPkDy/lKOwhbzSxuj6L95d/aIvf+feHE6VIQqBCcJjKWr7yefV/rqhGoQ4oYkwPsFisVpKwnUKzkkSg ClUgngXz+UJbzF29teu3tgNVwf31sGB/X8zS/7gR+9vkQCPg1Cnp2CG/b8Gi5duhSQ9BLYKejxmsvlAv kDU8nQB/wSGdhATUnA9Z5IcX8I8HPbnT0SUAVY21dThRiA4epOYwVAA+9x+TPK9/1VVTBBwP9Bpa+A5A RKX50wVqsAyvtesKegnlbiBHDv3j24Pr5XBP2twXaRBHhO4P6h9dCX2VQGfipdBb9EAGGVAK0KRAgFFv echqGh06EBT224b4te5jIRQKUawAeDRiEBmuGnLb88+88JAhWAC/yUBJb+4tNI4NeagXaMwO7sETUrQN 4VGBweUvrqhAZUExkoNcCVhUACmhEy+enKe8a+lUKhAHdOdVe2MtfPSowAEPz37t2DqYEObwUgEcH0W2 vg9r5V0JfZpIPfRgS+uYAaugQX95yGa388agW6BHhqUwMdcq97paLAQcFBGpYE+Ht5v79J6kdN+tHPXD Z6TT3v4/m1dmBSCly/7VeB8AClfxlgBiJgqoDfPzjZDfpYMjpwpEoElAC0pOBYGn668p5YtgYAON/fqk KGelmJE0CxOO2M800dgQj+uwdWw/Rba2Bq71LY+HyvHgL4gj9MPsCjS3D9tl/B95/kjaDHNlJquAUV7f 6h9dXPxOYWixLg3j+KEqAHf/D5fhgCmBJ+tM5vq/nTuB/3KqTGChq50Jq/tQRo6QfY+HyvmCdQ2X5xnF g5XyCFArRvQAM+HzoyllbgR6Ofdfu7CwESKJVKMNy3RX3tP4sXkoJeqJU4Afx05T2v0h+/Tr+1Rnl+vE 4NdCgSsCoAExmY8gG0KuChAtZv+5XVu1PwI9ipl7p/aD3cP7QestlstamFdhsaVAD1phL4fciA7+7j9/ TrTF7fd5cffh4Ff2qsoAwVgW8IEGgIeqk8/BObgRDI9PctEkElvn94pjdABCocuLoT9ElD7EqqEJwEHp 7pFUlgfHxcI496mEGQCAEg+GdmZuDzc/vh7oHVzhAAAU8NwT+1dync3rcKbuxaATd2rYDz/a1uEhBsd/ YIbHy+Vw2h/Pzcfu16+PBh2J09Uk44GkgAwW8CPvfweOWmpL+wwQX/yNHzISBc4OdEwGW+aWMPHzCK7w sj+3nMTwEvgd9GBCLwaZMVactW5LlhL/RlNsGdU8EqQFV2y2qA5wh0JcAnDRH7ZnU5sfdBkEQenunV+l fu3bsHk5OT5aE336yum92GiRMAj/+5l+fgl0A//dYaBX60mzuaYWXbNiMJ7M4egcOHD8Pn5/bDT1feg9 JXJ+CnK+9piTp8TJ/Hr8PrtT8ehdHRURh44x3YnT0Cl/NbIwGeG46yVqBnYQA23iAQTN7fRQq2/fy2Md +UBHxjfpPXp49NJMCJwEoCwtZgpeLWvQbDfVvkJiBa9uPxPZPxuhIwEAC8WC4XVt7L8wEPz/Rq+1lmZm bKYQDZbGSrHMzGSpwApt9aE5D7JuCj3di1AqYGOgKgRytsXwg3dzTD9Z5lsLJtGyxbtBk2Pl8upX1+br +KzVGm01idApyTgUQS9HUKehfYqdQXvT8OwOT97ZXHCEr8Ohq7+4QBtr38rnP96PfBfIAJ/EhUHPAc9C 7wU1v6i0/9wM/zJ5Xf4cq2bbCybRvcGmxX4QAlBGOMT9SARgQc9Nq4sfKVqwAkgJ+uvKe1sw8dy5d3JP 69SakA3FcwF0QQOwFQ8BeL06LX57E9eni8Tu1dKgKeXq/3LIPrPctgaqDDKMm//yQP33+SDyTokBBMxM BJAsEvgdoGdMnuHlhd9v5IANTwj7uSPKPn8UkAM5kL4GHCAK4EsAXZRkZc9ocBvzJpUxUfxkI3aJENQt RGMpbSX2CiMEnyETA/ONld9vR82Ki6lo1XBdDonpZLly6VT0mClFIBxeK0OARnNlaiBHD7uwsK4DZvT+ N7CfgU9Pgc9ozzpJvNkAzoPc/gQ2FMJIgwILeZ5v0rtvQXn0Kqf0zruafglyR1VALwIQfb57q+d1Svzx WANf6n8p+WZ2koVVECqQ17IZ1OKzUgZ/xpnF+5n9DnBcJYWnl8PmZMs4k2UQlgS/vk5CSc729VQ0keDZ dHkPEhOE8EAXx+bn9ob4/xfWH7Qrixa4UCPfX4VH7zOBytVmKgBOEb1/tYLpcTwY9yn5/Bh6AISwIuzx 6FAHzBXwsBLBu9Jst/An6lArgCYFuFFQmsew1Wtm1TOwgDswFMcX4lWafi+w+6wAp+RgJUDeDGtkKhUM 0DVFQDqoCwU7DiWLETAP7wxeI0fH5uf2hvT+N7CvybO5q9AOYihrDk8PBMbySg3z2wWrynwOdmA7+P97 WpAB+CqIUAYpH9Du8fkP88F4CJQBMRNL8MfZlNYjNQOUEoxPkkyafmBxiAjxOIMSfAk4M/XXkPJr74Fo aO5dXn/PuHFVYVkDQJxEoA1PtjxrMWb8+9fi3GE3e+5BDF+0vgv3tgNVzOb9W8vgv8pjp6WCLwVQhhCc AE+qjgT40VylUFm/SXFAAlAAQ/kgELB1a2bYN0Og13TskDQqqxfjXO55n+wFkDDPxoUn/ApUuX4NKlS3 BrsF29F1UA5gmKxelZUwGJEUCxOA3n+1vh7oHVcHNHs9HbU9Bz8Eten4IrTjMRgw/Qpav03O7sESX1Md 6nsT73/FHlf5TwwJXYcwE/DvCnxgrlioJL+psUgIMEMBRINb8M6XRaCwlkEtDBrTL9lVBAAj2fPMyTgt 9/kodLly7B+f7W6udXVADOwiwW5SlYSazECSCst7/Y3gTXe5bBjV0rAoCyAS4JUvAFuw38aDPvrixv52 Xg5wk/EzAkL1sr+HmisVbZH5dpA0EN2X+jCnCRAMsLTA10yN2AlQnC3MsjCYijxh0kgLMLLl26BLlcTv tsGCurktlWAYkRwOTkJIxkusTyHQf+xfYmKGxfCBfbm+Dmjma4vW9VAECS+XjdOJSDSdL7/IxomP2nkt 8H+JQAfLxvUpak1xdVAJX+OFPRJv8lAuCbvAQlgCSAIQEnAUnqY3zvAr4igG9Wa0rg+vheGDqW1wmm0l XI+waSJoFYCYAmAIvFabjY3mT09ujp6fXmjmaYeXW5ChvuHlgNM68uh/uH1hsf+xBELcrBF/wuMvj98c 0qy08JwBcU9Qr+uA1/L5gMxB2Qy0avyTMZXeCXCABnRG7Yqw6PwZ16LhLQ4vuJNvNhI8KhI3SS0LU/Ho W710gYUQkDro/vVeXCYlGfhZnEil0BoPxHBcDBj4D/8rlFTvBjDsCXDOjj+4fWq8dRlYMPsF2v479puG 8LZLNZscEnrJmALxGB62tsj02ePmnw80NJl41eK88TeGlQBr1EAtwcOYFU88uqt8REAlrMX5H2CuwG4G tHkBEVcOdUN9y9VkkyIqlMlHsPpGG4SW0cio0AePw/OTmpSXvu7b98bpEiAQT/zKvLNfDbrpJSwGst9m N2gfZ5YY2+D+9v71sF51qb4FxrE2QymWqbb42AMeUFXIqhFvAnZVIylJJAYBOQj/x/abBcVeAkQCdFUy LYsFftKpQSg2J8P9FmBr9w8AiSwMMzvUoBKAKovE6H4RaLyaqAxAigUCjASKYLvu5cIoIfrzd2rVDApy QggVMiCheY6TWKhQE6v9J7JIBzrU2wftuvYiEAExn4eG8X8GcD9C7wUxJQeQGT9OfgfyGv2pcDCkAige aXVTs27Rq0kgA5UMTXaJOQFkJUPhuTgdg0VCw+xgTAPT0F/9edSxSQf8wugBu7VqgrBbtEEDOvLofb+1 Zp1zBE4EMM9HOjgB/fj899tiAFe9YtUSSwYc+vEwOUC/Sur5kN0IcBPx1FplUHLCSwqPe3KuGa2jESDA d4LuAl/RQiKSmojhUTknxO4NPDR7AqMBFUD0gQtHW4WEyuIpAoAZxrDXr9i08tgOs9y+DH7AIFftM9gp 6Cn3YWxkECPgpA8uwc4FIYgvc0BOAkELcSCEMQc21hwM87G5f+4lOj5091vi2OMVPlRdpCvGNEPJIsm8 2aScBT8nNyUCQw0VZVAIYw4fNz+2FychIK164kqgISVwDKnloAXz63CArbF4pANz2HVw58iQDCEEHUcE C6cjLgP8uNXSsCCmA2lEA9mw/wTeBXIcHrHxlj/zBnGPATilENZLNZuTIwEQSu0+uTe0wImj5HhQGzoA ISIYDJyUmdACrgR5BjYxCX/iYFQDcTxUECPkY/RwI4zz+YXsP3cuDvWbcEPluQgnOtTeUJRHUAynoAvw n4JhJQEp+AH6cS+Q4zVROLqRKohAI4XERsFArj9Q1Hj5nyBNg2nHQyMDECmJychHOtZfB/3bkEpvYuhZ lXl6sr3RsgEQFepS3ENkuKBGxAl8iCksaP2QUB74/gny9KgO5vCCP5bduX6VmFGOfbPH8YEqDzGayVAU +v7yIAThqPhjfCtT8e1VTAY0UAqAAQ6Hw3IJIAXrkKwE1E0rgwHxKIyyRvzhOEJqVAH3MFwO1ie1N5SG gdgHWuvL4Eftoy7SSBSuY/yjhzflIxbUfesOfX9qQgA74V9MyM7x0r71WY+OLbRMOARAlgau9SBf7b+1 ape64GuBKYeXW5OCTUhwTiiPVdRGDz/NLX3dzR7AT/udZyojSbzc5JYnCuvL5Pws8H/KgCfMDPSSBwgh FVARUSMOYDXECOSAKPPuiCO6e6y2EAOR8zbhWQGAFcH9+rAZOCX1IBSAI28LtIII66v03+u4AuvdeUA7 j41AKNAKiNZLpq6hSsB/P1+FHAj8CXSCDKQSYSCWjjySuhwO+PbzaGAr5gd5EANbp9OKkwIHYCKBan4f tP8gqUElAxFJDyAi7wSyRAQR8nCZjkvgRwm1KQCMAEfiyZjmS64NkDHz5WRFCrx3eBnwPfBX7J65vif6 4A1JUeTOLoD/AG/ESb19dhHiDJZGDsBPDTlfe0OYCcBExKAB/jFGFfErCBPy4SsOUBfEhh5tXlSuaj55 eAz1unC9sXQi6Xg/868LGyeiWDMB7fp8TnI/tdnl8CPx1h7jrIVAsDKqFALpezhgLewKf3Jqv0AyQZBs S3GahyCnAYyc6Tgxz8NiIwgT9p4NPnTf0Akor48rlFgZifg59vjcYNVIXtC6GwfaFqq6af1ZfZBK+k10 Imk1Gmpg7Xkcd3ef0w0j8K+MPIf36QqTafwFAVgLG0H/j51aYAPuiCqYGORMOAWBXAnVPd3nE7VwIU8B z4/DEFP5qNBJIgBBfoJQVgM75pamrvUm1+Ahq+jiSC5cXNa6rW0vKMsnoAfi2gDwN436SfE/xCHoCGAa YuQSPIw17xfqIN7pzqhvHxcWNTUK0rNgL4/pO8U7qbSIASACcB6TNdcj+OEMBU6gsDemqF7QuN0h/BT6 8S+PlYNdpUtGfdEgX+zWuaILWwpWyVGQRRcwkU7FGkvqmrr5ZkX5ikny/4TSQQWQVMtJmvaB5k8Gh4I4 yPjycWBsRCAPQEoOm31sDUQEcoEuCAt5EA9fpJk4Dk6aNuOUZwmyQ/lfsU+HSMGu2gxBLrxXadAJAEFA G8NCgCl4I6LrBH8fq+ib9aYn5b7d8G/EA1gM0oNLUJi2A3Gf066Z7kAS5duqSa7OqKAKj3D6MCMAGIoO fGP88m+5NKAMZlN3atcMb7PuCn5VLcI1HYvtBIABRgtQK6VvAvP/HnKphe/0gEO4I1DvCHjvt9SaCyLV kqCyoVMNEWjggkIw1BNA9QLFZnBs49AfzrBwVU9PwuJYA1e/wDNgGf3nPP7xP/xxUCRJX9JgLgnp9Lf5 vnx8+iV1QYSACphS2wZ90SyOVymoyebcBz8lGZdXLwpwZKNg3Yp9SXlOdXP+vrH8HQsTxczm8tJ1eJCn i646h8luBEW2z26IMuuJzfWg0DYs4D1EwA6P19Yvfpt8qn/lLwmrw//Rzu+R9HBXB73ypjvC/F+hT8tE PSdL25oxnO97fC0LF8+YDTf/0An5/br4GES+84AO8CPyUBra4unfwjGAI6Kc/vQwS7s0fURN+BN97Rzi gQ9wl80GUG9dWd4UkgwTxAbQTwrx9gaqBDAy0+lpSAJOFtBCCBX3ps8vym1+fKJM/vkvzUy0v3t/etUm caPvrbNc1+NvgnbRIxelIO4FqB75vgC8TSJqMTgS3tvbV4/jDyP/X6RxrAf0GxQZUAACAASURBVH98sw oDMpkMPDzTq4UAVhVwdadGApwsxMdjaUUANAyYcwKg3t8Vv0senEp9yWio4JL9LhKYa4Vwe98q8RAUG/ hNHv/2vlVw51R3+fTif/2gAE/vh47lxVOIpARcLfKey3xbQk9rqvEEv+/23jCePyz4U69/BJfzW7Wpvk gCi3tOw63B9kAewKoCYEgDO7/y+wcnu7VEYJz9ADURwJ1T3V4EYAKxS/674n6X1Qv4qdeWYn5TvM/Bf2 uwHUpfnQgAXiOBf/wOhoZOiycQ+5KA9HytTTzLT/xZjPONoCcHgizq/a0WCkj3ccT8mmV+E8gDPPqgSx 0M8uBkd/mg1xfy0JfZFMgF4MCPAMCv7iwTwNWd6vNsZIBfczm/Fc6e/TT2PEBkArj93QWnfHeB2PbeWs HvIoHZ9v6cACjoJc9P90lIEp+CHu32dxeq8SnuZw9JAqYrB54v6KUwIAB8wylA/Ipz/mqp84cmggoZDL zxThnYlVN8EOjZbBZSL+RVY5B2kKgAaE4AFORGIqhUAmgeYM4JwCb/fT24b9wvgT/u4R+zRQKuTD8Sxf Rba+D7T/Lwz+KFANAlEjh8+HB5uGX/mBpyyUkAwwEEtCTfpStvhrGV8nzq+dK5f+IRYPQcABwJTioHLs 8fBfRU9uP9ot7fwoY9v9aGgND5/qkX8vLBIhKYkQBgSIGbkgA37AW4c6obRkdHY08ERiKAUqlUE/ht8b 8t7kcCoOCvZyKgzUO3962CuwdWO5N99w+th5+uvAf/LF5Q4JdIAMF/+7sLkMlkynPtXxqsgp8qgf6xMg gcuQDp8dJffBoAP7+iZ7YBn96bDv00nv/HxoFTJRBXvG8kgsqxZIt6f6ufFDTRpoB+Ob8VXkmvFfMAAS 9PCECpAPK1JntwsjuRRGAkArj93QUoHe+KJN9dBOADftMIsHoiA945iCTAQa8k/0AHlL46Abe/KwMfrz YS+Pzc/sDBl4oEhBCAl9Q4WCVZn3r9I+/SHSUCUxsvEopGUgz82vl/0lkArH8grnIf9fg0D4AkoLL7eE oQkfuvpNcG9gdwYMNYOkAAGlEYiACPGUcFEGciMDQBlEolq/z3Ab8p/g8LfhsZzKVx0NMzBabfWqN5f/ T4UBiD299dUFcEv0QA1OvTc+9tCsCVB7DF7GHq9ijJMelHgU8VhdPrU9DzE4Eyv4nV81PZT0FPQ59Fvb 8t1/wrR4PjgZ4I9luD7SoMoKcBcxBrBABD1ecNhu/FSoCpIWhWFYAp+z/zqjl2dxGAS/q7wM9HguHPMt te3wZ+SgIY40NhTAQ/JwEkgM/P7VfApwSgKQAhEUjjZqknwFi68y3bsZKd8vKVPAElnkC8jyCnQJc8Pz WDCgjr+aWYXyMBPJk48xu4nN8KD052VwmAJQSnBjoCPQEYw2sE8I/feROAIoJKqKEqAZVE4E9X3qtpTm BoAnj0t2vW2J0CzzcB6JMwtM0V4OCfC+C7PD8/TPT+ofXw8Ewv3P7ugor58cpVAN7ncjkR/KL39wwBXE rAC/jksdYvX7nS75t6/aMA4PmR314HgBICiDvm554fiQB7///9wwqQVIAUAgSMK4BKGOBjvBJQKBTU39 CsEcBPV94zJv0oKLDlVyICTgAu5eAzXEQC/2wQgS/47x9aD9NvrQkcP1463iWGAJQEfrryHqTT6QD4rd 6/RvDjawFAO+r2YkaftPTSzTSi3Pe1SrgRR8yvPD5/bu1BSHW+DYt7TsPintPQl9kUJACmAgIhALtyAi h9dUIPERwKYHx8vDwfoHJ24OX8Vnh4pnf2CODhmV4r+LF+bVMBdAOQy/vT8WKcBCgR8O85G57f1+NLwE e7c6ob7h9aL4YCUBhTkr+l5ZmawC/11PMEnZS8axr4gxP8UgKPx/ZiUo/H9mHMAv4opT5J+qfWHoRUql 8ZHhKiKgGCCpB2B+J9OXwYCpqNMAgBSIlALBFGDQNCE4CUuOPgl5SAqQLgI/1tW4tp6DHbXp/H/JwIqN yXrHS8S5HAg5PdCvSFa1cACmMw3LelDPTFnWbwW4DPvT9efYBPr6qDzxP8xho+v4b1/J1vq65Ak+yvJe anjxf3nA4QQF9mUxnEE20aAfz7700i4DF5R58rfXXCSADS+zmhoALAMACfLxajJQNDEQCP/xFwONmHA5 +CUlIArsSfDfyuuYBJe34EOgX87X2rNNBLnv/WYHvA++P14ZleFQ6MZLrMnv+lwVBxfxjvb7rHpJ7k4W 2ZfGs5LwTw1WcQkGveO6wR2c8z/6gAFvecLn/vCgFkMplqgw9VAKwvwAZiUQFc3Wl9H1UAQ8fyMD4+Dh NffAvX/nhUfQ3mAcKSQCgCoPE/lf68BCcpAQnYtXh/02iwJL0+nQ4kyX6Xx+de//6h9fDgZLemAn668h 6MZLogtbDFHvOHBL9EAhLgTc8b5b3J29cKegp8vK8ogNSOEe05DAl8pb+U7V/cc7oq/TO/qXr+tQfL1v wybHy+10gAtDuQyvcAiAUCwDyATTlIeYDPz+1Xr3//SV7rCfAlgVAEQCf/SOU5qgQ4GXACkJSBpBIQ7P Sev382CMAn3neBn3v8O6e6oXS8S4H/zqlumBrocIPfQ/qbPL8N+CYiSL3+kSzvLTI/Mug5+BHk0mP+Wo g6v3qOSHwu+VNrD6rcx+Ke07A7e6QKbq4AWDLQGNdLCgD6vRQAEgDmAXCH4sMzvXA5vzVwdJgPCYQigD unKn+kJMPv05TD8wC+3p/vKsR7qeSYZKzP43wb6LnsvzXYrrw+gh09PwX+g5PVeynjHzbm9/H8vkZLd8 a4nt/bnvMBvYkEKODpa5XwIEzMv7jntA54Zot7TmvJz93ZI3BrsB1gLA3//ntTuRsQuwIrYYDJcz8a3l ipAPSLJGBTDvQzqQJAAoCrO+HOqW6tNdiXBEIRwK3Bdg3MrqYcqghsgLd5f4kEOFkkQQImjy/JfluWn0 t+yevz6/n+Vu8uP3V4RULg5517opevJZsfBvScAHaM6KFBmJIfXjvfhtTagzIZrD1Y/jpCAAhMJABq2P RjjP/H0oAeX7dgHoACn37OnVPdMDR0GsbHx5X3xx4BbA0OQwLeBPDob9fgzqluBbypgQ5tpr+LDMIQgG 06MI/7kwC/tIknLPAR8NTru8BPVUCYLj8O/rgJoGngD7V5dp/43gf0npLf5PlNnX5a3Z8SQOfbigAW95 yuJgEtBMA9t0YASgEIViEAmxLAx6OjozA6Oqo+E0vzfIOQT1LQmwBuf3dBxd6YqOLlOBsR+IJ/5tXlGu DpPZf+cYK/1oYe7vFNnl+S/PT+4ZleeHCyG07k/iNSzB83+KkKqDm2N3n5MCSwYW/13hXv2zw/zQVI4C feH9WPSwHwDUKBngAPArAlANGGjuVVAhBHjD0a3iieH+iaF+BNAN9/koeHZ3qVpJAm/tpIwDcMoE1CXA Hwz4jb6/vE+r6JPh7vu4D/4GQZ/EgCtwbbreCn0p/3w1MC4ERge2x6r2q4ieL1bRI/rNdnJGCU/Qaw89 19XAGo2j+GBCj/KyEAPRHo3z+sKIOeXdWuPrEE+GIF8Pxavse/BykMoCpg6Fh5JPnDM706AZDTg3goUD MBPPrbNXVvOgOQkwCv1YeV/67NQknG+1FifZrwM8l9k9eXrul0Wiz3aXH/6x85wR/XVSUDo0j7WkBPvT 69l8IAWt/3JAObqUTgC3ldAVQGgyirkAANA7D7r9oF+CIDv04EXn0EY2kY7tuiJQBxP4HpFOFYCABXqV SCa3886jzKGwkgbALQd8BIXJKfSv84vD6P+U2xPoKc3mNS59HwRnh4phf6MpsCCsAG/jhJIPX6R+JjDA WcwI8T9PQxXte9piUDnbv6JM/PHvPXeAiQzWbNBFAxmgjEK5JAFfQyEfjmAH5/fLMaR47hw8MzvdomIa 4CYiWAy/mtgbq8iQSiZP9t24xpEjDOJF8Ur2+K9X3kPpZUqdfn91MDHXb5b/D+tZCB7bPQFAkkKO9F0H Pwb9hbluwb9vp5einxZwI+k/9r0r/UjgY3EgDbF0BlfZAAmFVALuUAeC/ArcH2al9BJQTgw0KKxelkCI CfAGQyV7KPXyX5b1IRcQFf8v42r4+tvCap75L9FOT4GD0+BT8+piGA2p5KwG9SAVHNlwisJBCHvHd5fg r+tQe1OYVWz+8DfOb9MfE53LdF/X8h2L0JYKLNSQCPPuhSZT1TNQFLflpZkRMAmxk4MzMTHwHwg0BNRG CT/jOvBncNmgggrtg/TJY/akefzdsj4E0xv8kymUwwBHj9I2sIMFumcgJRPL0N/CbQS1f8mrCJvxfy5f e/kNfjfer98blKAxTuBnz0QZcCv3TlOwAxW19uGHqxArsy6KvPvSiWAnkSUKwqVN7HZwVQFRALAZRKJT UO3EYCNulvAujU3qXeY8Likvx8So9vnM89P15tdX3u/dHrS9Kf2nDfFqP8T0oF+IJfIwGJCOhzNk8vSX 307hzoeI+vr3utSgidbweGeFDPrzb30Pcg4PH7CglAVAAqBLAQAN0TgAB+cLK7MkPgRZEE1H1FPUihAF UC+nCRftUJyIeGJkIAP115z9qo4yP96cx7evUdMVZrhj9MDz8C3qebzxTjc/BTsGPiz2Tn+1utOYDZAD 8nGvo91XNsg45SKr5e3gV+IvcV8NcerL5OQS15fgp8CfzsOQp+JAB1DmCFAKgp728IAbReAQX+lH6PQ0 KxaiD0FGivky7CqYEOfatw3ARQKpXg3r17zsNAXVl/vlOQPo7T+/t4fZ8kH5b2+N59W4Zf8vgmzy8ZZn W1PADKf2wCYmGAFZwWpeD79dJr/F611vLHNm/PwY6vIagpCSB4ufdHw+e43DeBf8dI9bnml3XZX7nSFm gVe0+0BQiAEgEnAH3zUBX4uhpIiRUEPhBEUwWEAOgmISkPEBsB8IlAlAx8vD+CE4FPy3E+Q0Kjen1KAg h6HvPj8zd2rYCbO5qVfd25BL7uXALXe5ape2r4/PWeZcq+7lwCUwMdcHNHM0wNdGiKwNdQ2mkKYMdIVQ WwGrgNxLVcvcHP7aXB8Nl87u0psE3AlwwJRyIJmgNA8OPPREmAyP/FPaerp/9YCEBSAHyAiGT//nuTFg JwFcD3Bmj7CioKYHR0VMwDxEoAEvh9pvsgeCng6T3PAYTNIZjq+nwfP71Su7mjGb58bhFcfGoBXHxqAX z53CL4bEFKu8fX6RXvv+5coj33decSONfaBC0tz0BqYYtme9YtgRO5/4DL+a2KFKjXf3imF0pfnVD3tB eAVgIoCXAwmoDqA3YfoDvBjxt1bGAXsvki2CXg+xKBzZpfrn7WC3lILe4UFQCSgJLhHgSggZiPEDORAM sf2EqB1dmCZQIY7tuiEQA9PKRYnK6dALCjyAT+mzuanQSAm4foTkGqBnznBPqoAN+vQw+PAKdXn+cR/P j4684lsGfdEg3wLS3PQGpxp35lr49kukQygKs7xWYgUxtsWCCHBraPcfA7SngKzFJyLw6gc8Czq+r2W9 hiBD8SAIyly4CmJMAIgSuAatwfjgBMewMefdBVGS1WbiPGXBHuEhwfH49XAaD3n5mZEcE/NdAB13uWeT X6UO9PVQAlAFcJ0dZLEMau9yyDi08tsIJeeoxenj/3decS2LymqQzsCtBxph/O9aPP4T0lhM1rmuB8f6 sC//ef5I0EUNNIrKQs8xu5fMeTetLVdh8H4E0kgFJ/YYuS+zz+3509AqXj5Lw+BD4ngom2oOemMwMcBO D0/Dz+ZwrAdHpQLARgOhEYz7r3IQDTTkGaA3B9ji8R2EiBA5iDnHt2Ku3pc/Rr0fMrUAug53P++GtUIa AqCOwHwNZXQy5gTs1Uu+dy3iXx4/L2NuDjfefb5d/rwhZIbdgb8P6Le05X24CxAkDBzwhBA+pEmzY4RB okosiB9BDYdgXquwrLQ0X7MptEBVAsxhACuAjgswUpL9Dy3YK0ZwAVQNj24VrBL8l4H08vPaYSX+3nF0 BvmvQbmAKEYYK0H4D0v9cNCewYMcv6JD29r8fn980vw+X81nKzVeY31d83K/9hF6BK6DmMAlZtFLKYIg dHDkAlAWn8D/1qgxDmAHBoaGwKAOW/RAAoo8MSgDTWO6z3t11t4KcApgk8Dmzu+flj+tzF9ibl/ak3d4 HcSgCmAaD1Gga8kA96eA5++lxSHt8Q61PgUwJ4NLyx3HCFBEDmHmDzkNp95wI/G+7pAr/WG1BRALZ5AN X6f1UBPBreCLlcDoaGTscfAtD4//r43sC5gJj59gEuBb802TfK7sEwZHC9Z5kGcAns0mPu7fkV5b9pjL fpOC/ba9owED4TgBzLPdcEoH1/KvWl+D4u7y+B3cPbSyRwOb+1HNsPb4Tz/a1lFdf8sj74pPPtcgkQ43 8HASjAftDlTQBYBpR6/2lY8PBMrwZ+gH4oHe+Cw4cPG0OAWAmAZ/5RNvsSgGnPAFYF4gK/VBWQMvZhwW 66vpJe6/T8CGYR4DbDQRwWBTCbRGA8Ugtr6yYvH1dSz8fL24igYssWVbfVonc9399aTtqSrc5PdxxV8h 9JwEgG1HN/s7oKcJ9QYKItUPuX+/9R/le7ACkBJKoAuPdHwEQlAArmMN7fBX5OBBywPjE9lflSAxB9Pp 1O20/wiWJU+rNeAFcYYByN5fk4AG7Tc/SeNt4kUb7zIQEL4E0hACoAtFfSa7W25kwmU/a8DPgiGRDw8m Eh4pUQBC0hmpQA9/5wdSf8/vjmoAKonB1YLNaoALD+PzMzo/Yh3xpshxu7VsBnC1KxEcCP2QWJ5AB+zC 5QTT6S1HdJe9M97wTcvKbJPcdfur6Qt78uyH8eAnAV4AVm8nxY8Afeh6/R7rrZAH8NwEej5/2hCrg12F 7+/6sQwHDflnJsT0wiAIz/YSxdlf8+JFC58goC7wcol//oVOGyAhjJlEMAWgaMNQSYmZmBYnFaseWtwX YVQ3+2IOVVAvQhgCjmQwK1enqXXe9ZBq+k14YDP0p612NuAgGE8tKmkVkeX2/cb195bXHP6dmr3dcIfD Tc4cd7+EcyXUoFXM5vDQBfIgNtHqAwKyBABuwe32/cBaidK1AdJtqX2QRDx/JaCBALAVD5TxXA1ECHAv 9nC1JeXYCUALDbL2zMHwX8P2YXOD17LeBHAuCz/EXwm0BNx2zz5wzgxw0vgUEXgoc2kYR1XHbIq/o5Zk vm1wh8TPRpJ/4SCX7nVDc83XEUUhv2qoSey6jnNoLfQgiagmC7AHn3H5IAbhkfOpYX9wIUi9PRJwJRAk AFQEt/VAH4xORxEoBvDgA35vhI+ah2c0czzLy6vJwHoEk+H/D7Go6nMigAkyxP+nHgyjfgYDItTELPJf FrBD4lAFQAuGGHqwBsAEJZ70UAH3SFIwCcJmToADRl/3GOYDabVSEAJYCatwObFAAF/2cLUlDYvtCZgc cWYLp9OCoBhMkBmGL7uOx6zzJFAKgCAjE+Bz8Ds+nobRPwqce1KYBEr6Z73GJb2XWnfj4kAd8SXgKA16 wysESd91chANrI8+Bkt1YlcJIAAneiLRYCCGb/mfyHfrg12A4Db7yjegCkeQCxKIDJyUm4nN8K9w+t18 CPIYCp8Ybu98cegDAtv1HIgBNAnGCXrvhvm3l1eXWSr+D5tbPmolgFYE7w42u1PrbM1DPN2DNO3e182z +uTwLwFVu2aLPy/ot7TgdO/P3335uC8pt7ZfT2/Mqy/6aZgbZhoqa9AJL3xz0AvALAS4CxhgCYOJMUAA e8RAQ+E3/iUAA/Zhdoe/njBj293ti1QptBkMlkAt5fgZ/Mm+Pz50yAl95jVAISSA3e2vW69pxEMI6fQw N/WNAnSAKpzrdVo4+a9U+P/J5o8yIB9T6yRx/lPwW6NDaME0FgD4Eo/80VAFMXYOwKgMt/kwKQiODHbH Xuf5QcACUM3wRgWKC7wE4HfuA9fj86d2Dj8726AuDAdhn9OsN7TEpAUgaLe057vy56dIk0hMeB57j0ny VPr2Q+gh2nD5HdfqnOt2FN+pfwaHgjlI4TAmBbejkBaJ6Ze2w2IswGfG2UGCMdnB4VbP4hRFDZBGTqAo xNARSL00oBcPAjAdCTgKUrArfeFICPh5dAT42CH4eQTL+1pkwCXAH4EoDFJC9rG3Ft9PSm91m+xudeMx /QJ0QEdFcfN6oCFNDI1lwTAUgxOpf/NhIwXlnST1MctPOP7QKkCUBTD0CsCkAigOs9y8S9/vRxnATgmw OQCIADPCroOQHwcwY0EsD95nitkQB8AW4DaGhPHtU27E3WyxtifC73bSSAnYDa/vwJcxhgJIWJNuukID 48lBIA7wCUt/7qJDA10KFtAsL4H3cCxkYAxeI0TE5Owufn9osE8HXnEiPw+dmANAmYhAKQSoAu7+4Dcp Pd3NEcyAFQEri9bxVks1l1hJV26GTMROAkhwiv12qLe07PSlxPgY/36rnKYZ+iAuh8W1QAvDHIRxG4wG 8kgYk2Efzln4mCXt8DcL6/tVwBEHoAaPwfmQDwzUgA4+PjRgVAR3zZTgamgz+S7ANAAvCR9mHAbrpKg0 gpIeRyuURVQF1bAsAPeHoHCXACeLrjqEYApeNdgak9cXt/o03ILcCm2j82AJ3I/YdWAjQlAGMLAa798a hIAF8+t8gIfk4AtBIQRgFIX+ubA6jVy7vAz5OAJjIY7tuik8AcKALNdozA+m2/glwuBwNvvAPDfVvUyb PDfVsgl8tBLpeD3dkjNX0PVQJMAPgc6KYrDQW0xCBRAOXDO6KFAZHBX9kCzON/fd+/Dn6M/zOZjFcJsF QqRScAbAKanJyES5cuOQnANPGHWi0hQJgmoFqAbgI7dv7ZFAC/0gnFI5kuWL/tV0YCSFwd7BiBXC4Hw3 1bAseg4fkH9FQjPJugdLwLLue3Qi6Xg/XbfiUTlqG8qfYIxOzpQ5FBBfTU8y/uOQ2ptQc1AtAGdPgSgH BQSBiTyox68i9IBA9OdqsEoG0bcE0EwEOAQqEgEsDFp/RhH/xqIoCpvUsTrwLU4tldV2o2+U+fn35rDU ztXSqSQGJkUPH0CHq0uwdWB8DO7ftP8uqKo8ofnOyG4b4t1X9D5jfOK52yE0UR+Hr6ADmk0oE8gEkBaP v32Ygup/efaDNOCXbJf+79cSCsrAD0+N+nAlCTAuAEIPUBfLYgFeqUYMwD1FIF8FEAkjc3eXqTZzeBnj 92yX+uBDA5qCWmJOBwIvAhB+KJOfDR46Onp0eXUaPPIQmUvjpR3pBSGIPvPymXz4b7tgSUjPZveCFfk7 T3ie8DYBeuFPhPdxwtb/TpfBue7jgqKgA0b++PoJeIQCKEynO0hVg7+8/i/bX437IJKPYQoFAowLnWJp EAbuxa4U0AmAdIYjcgrQRQsEqADuPhXebKAfCDSvAeD/8MQwTW5yrg3/h8L1zObxVPPKYAN11NaqD01Q n46cp7AIUx+GfxAjw80wtTAx2wO3tEzm/UKPsje/zKlScBUQkg+LPZrCoDSoM66Iw/MfM/0SZPB/a5fr NaTvwF+v7ZtdIA5LMHwJYA9CIAmgQsFAowkukKRQAmme9zAIhPMtBFAmEBb7r3IQCT/DcdUYZfM/3WGu jLbFKSWv2x0um0HjIbd7dNDXQEjj7jHl8CvA389LQiKIxpJFD66gSUjneV90FUEm5a+29IWe/9nMHjL1 u0WTMq+TUSWHtQdQJqCoBcrXkA14hw13VCzv6LXX8Vw991HAnASAQw3LfFWArk4MfJQRKQa9kOHDYRGE XSR1UAkuyX1IB0YOnUQAcM922BNelfBoBvA/3u7JFAUo96fBrn+3h+mxpAFcCVABTG4MHJ7vIgDUfyLo rEN4FfeqyBn4BdUgB9mU06AbAhHaa2YJ8JwZpJJCBt/LHF/pUNQDT+HzqWh/HxcTh79lOxBTgWAigWy6 3AhUIBRkdHjZUADn5UAFKsb0oQxpUDwPs4wB1VAVCQS2rAZkgGw31bYHf2CGx8vhd2Z4/A7uwR6MtsUu U623HnEuglw6+lx6FLpEGVwPef5AEKYwCFMbj93QX4Z7FslARcnjysxHeBnX/NskWbA7E/JwOc968RgL BLT/T+YQhAMmPd3xz/PxouzynI5XKB+J+eB5iYArCVAinw8f56jzwuDKVvLSGAbyIwCcMThOnVpAJcnt 9m+PV4ejH18BzwPMHHQY7g5glBHysd7wqQAFcBSAQPTnaXh6OElPlhwe8iAqsCWHsQUmsPVvf7W4Z41O z9uZHkn1b6C7T9Brv/Hg1vNMf/Qgdg7ARQuHZFJACsBNCBH9NvrYHrPctEBYBhQC3e3wX+uEkAwY6fST +btwObSKAWMjB5eknuc7CbgO8iAf45rlAAieByfqtRrvsAnXtyHuvbQM/jfyn2x92Bd06Vf1/SNl4TAc Tp/bUTfwLgD8r/qYGOav2/Iv+jJgBDEUCxWA4BJr741lgJ4PLfpgCiEIBPAlAig7i8PP0sfk8VQBgSiM M4WDkxSEQRxvtTk8IBiQS+/ySveWkfErBl8qNaau1BVfaj5T/MASBhomeWhnfwk3md4L660/o6n/1X9f 4vVgCPV93wFOBcLmedAeAr/yMRQOHaFWMl4HrPMg38SAC2MGA2FIAvCZgAzz39zR3luj99DR+7EoFxen 8OaEkVmMBfOt4VWQlIKgD7A25/V04I3jnVLcbmUYEe2vOjchC8/9MdRyG19mB1GhDr5efbdTW5PiGfB6 Dur+4sG60OcOPjvsbSFdBTY0Rwdacq/2H8bzsHIDYCoGPBkQCGjuWNeQDp7ECbCghLAj7gl4gAgRsF9P y52/tWieB39QIkpQAksLuIQCIGG/iRNCQVQIng4ZleKB3vKicCBcD6lO6ien0FehI2iOCvyH+cCEwJQN quS6f/2MBfnipUBqvpBCFp8i/Ai5X2Y04CxCrxgT4REQAAIABJREFUfyaTiaUByIsAAIR24GtXjInAi0 8tEAnAdnRY1NmAPglAKSFoAvyNXSvUcyaA0+dR8k8NdGhVALo12hYOJJEHiAPsM++uDOQLpDAD73mOoH S8C6YGOsqHbKbSykyeXIrzowKfJwYR6FLyjyYAS8fNvfzasA7LgSDK+1eSdUaFwAeJkFmEZbCnRAJ4NL xR7f+31f+LRf/4PxoBVCoBpjzA1EBHqNODOYDjAL8tFDB5eh/vLsl9fi1sXwgX25vgxq4VsVYDXMAPA3 4fQsAeCt4+je+ZGuiAPeuWwEim7O3P97fC1N6l5WO1FraUz9cjBBAF2KG8PfkekufXwM9agF27+VynAV W9/xAADDlPDqIKgG9B1jYjoTKoeH8s/8UV/3sRgKkUaMoD3NzR7EUAEmjjVgAmApDkPQU6f4ykgGBGQC Nh8MfnWpvgXGsTXO9ZJnr+JJKAYRSBLSdwe98q66apc61N5RN08ShtyRZ3igQQFwnwsCHwfda9Bqm1B8 vbmEnMj9fU2oNaC7BrN594BgACGh9f3Qnwj98pAhBJgGf+v1mtgV+3FxUhoAKR4v9a5H9kAgiTB7jes0 xNDvaJ3WvNAfiQgSmRR4HOrxIRSK/fPbAaLrY3KRI419oEhe0LYwe/y/uHUQLo8W/u0OcnIOgL2xfCud YmO+gR+JQAYlQBUnyvWo4xl1AhhsU95XHfD05268DHwaBc/k+06f35HgQQAHnF+wMMaeQggR+vurcPGt 2ReOdUtzX+p4eAJEYAxWI1EeibB7i5o1mNDXN56CSrAKbKwP1D60Xw42P6+syry9VjBP39Q+u1HgB8nR IAJ4I4Q4GoYYAk9aWhKRfbPTx+CAKIlMlnMp96eR7X0/r+1ECHBnokg1UtA7r8R/Abrs4TgRD4//gdYL ee6dQg6v0pARiJYKINHp7phZFMV7X91xH/+4I/MgFMfPGtdx4AFcDFp8zgNLULRwkDwoQCkqe3efi7B1 Zb5//RK1cB1L58bpH6nDjVgO+Ve3sT+Dev8fD6NvAbCMBGAlbQcyOgltp7z/e3arIfbbhvi3YkuKYAhK sV/ET6wz9+pyoApjMDtZODDUZzAfg+lP9xxv9eBABgrgTYNgbxHADuGOSgRwuzMchXAfgqAhrHc6lv8/ z0PfQxVgFMBEANJXat3t8m/fHfI4Fdur/YHgL4UuyPFQALAWhJuzCg51YJBWicjxOMRjJdAZJIrT0IUw MdqlyJMv3utZRx847k0WXvPyQSQED6812HJiKYaNPkv9T/X0v8700ApkSgaWMQDQNQAdDhoRL4fAnAB/ xhQgHq0SkR0OQdB7evdLepAMkutpftes8yVZb0+T74893et0rlNwrbFxrPRDSBv7B9IexZtyQ86CMoAI kEIhsr7aF3T6f111a1DKjDPrUEIAW+YNJRYOWDRBj4//E7p/eHCXZuoIUI8L1S918t/f+xEYBPGIA5AE wQIjgRbHSa8GxUAUzGY3gb+KnZAOqrAjgJ8HskBpd9+dyigGEfBhJBYftCkQQK2xfWBnoJ/J4EEMaWLS q39/LEX2rtQcjlcnD/0HoF8I3P9wbk/9RARzD+D0EAijSo9KckQA4QDRz0YdpwxGYQoPfn2X/X/v+ZmZ lQ4I9EAMViNRF49uynznLgjV0rAoNDpLMD+PTgpPoAXCQQZ9fe/UPrQ6sADnwfIvjyuUUiAXzduQQutj cp8Ju8vpbhpzJeAr0v8BMAPwI9te61ACk83XFUA//UQIdKFKbWHoRUql95f6UAPDf1qBCANu8I4C99dU I8LFR5f2GjkYkQ6PvS6XQs8/9qJgApD2AMA9qbyoM/BzoC+QEJ/Lf3rQo1JDRuBYA/k82bR43PwxKABH 4EuS8JYCxPwS+RgJbo4wC2xfgm0Cft+RWgdWKYGuhQuyBRNpu8P5f/NiLgh39WJ/YMBUkAW4ClXv+JNn GjkWkGAe4QRPkfZ/tvaAIAMHcEjo+PW8MAJICLTy1QZjtDYLZzACYlEEcyDh+jl41KADYFYPL+FPxcAR i9vkQAHPAu8Mft9Umcr2X0172m+gFoVh/bk/sym9TXrWoZ0Cb/0Pq/T4dfYGIPzfqzBKB0ZiBtNDIdGq oRAJkO/Ep6beztv5EIwJYH6MtsMoYB5/tbAwSAdXVqYUaE+VYBkiIB3+w7vfqGAq4cgEQEJhKQcgCBTj 4G4JVt29ylvdkEPkv08bIe7uhD4ON9Op2GVKpfkQBVByj/eUefRACl40TOm+J+7ADkwCfSP7C/wEACdL /ArcF2lf33OQE4rPePTADFot4QNDo6Ko4Lv9jeBNNvrdHAf/GpBYGThOgZAr55gCQVgIkEwrbc1hIK+M b/PjmA6z3Lgtl9G5jnKMHnQwA8q48lP0oCpeOVkWTrXoNUqj9Q9+dlPRsJaEC2gF/F/4wETB2GxlOEyW dg8i8p+R+JAKQ8AIYBEglMDXQo0FMCuLmjWQ0QQfD7qoDZUAC0JBiGCFwEETX550oCSiHAnnVLNG/f0v KMEfQr27bNvcwPQQBPdxyFqYEODfRIAuf7W8vef91rsPH5Xt3zW2r6oqktu4a4n8T/Ac8f9sDQieqU4D unymPVjOU/Qf4nSgCUBPi+gPHxccjlcsamIAQ/vaIK8DlHYLZyAAh6mxqweXife9/SYBgFQMMGsXvPAW oF/uaX6xP8mP0nnh+7/STvr8aTr3stkPij8t/HYCxd9u428MMQPDzTG/T+YU8KIlOCRzJdkM1mA+2/Uv dflPJfzQRQLE5reQBTMlACP15NZwn4hAFJVAEkIsB+AKkKIHl+nzDh7oHVsGfdEjjX2qSuYWzPuiWwZ9 0Sd6uuw6Mj8OkQjnoGPm/l5abF/80vw0imq9rxFwH8peNd1T3+UtKPEIIo/RH4/CrZhMX7JyT/YyEAWg 0wJQO556cmEYArDHCBP4lQIExY4FIHlAQ2r2kyWmphS7h+fM/YnQJ/2aLN5cfNL9cX+AXpj518EvipLV u0We0GDIC/EgJQMpCIoTqnb0g3Afylr04Eh3wgAZhIgJOB4P3DZP9nhQAAquVAqRowdCwvJwOfWqCy0F KWOnCa0ECHdYxY0grABnxJDbiUgE0VjGS6/Hfb1dCYo3n7xZ3VZho8nKMewZ9Ka808NvDT+H9l2zatKY jH/5wMJEKAibYg+G3yn/YLUPBzEpCIYKJNhRDU+4uz/2LK/uMKTQAmFYA/4CvptSIBSODHZBXmAuhA0R u7VogkMBc5AAQ+JwEfMrDlDNBu7miukoAtC2+S9Y64XcX4qTSsbNume306o7/ewI8KoNLF5/L8uAGI7g fgCoATAYyltdeqY7roeXwWEsD2X9Lw421IADbvn6D8B4iBAHg14PDhw6IKMBHA151L1OYhTgLYMzDbVQ AKbltegF4pEdjAbiKK2/tW6eO0TKD2rctzAFdAzoGveurrEfzNLzvBzxOA5/tbZfBj8w8CnocFtN1XOp TDpACu7qwmAD06C437DIj3l5J/Sch/gAgEAGCvBoyPj8OedUvEXAANA2iX2sWnFqhQgBIAbiTiSoCCNe kcgOT9w5CAb48ADQlCgdwHsIaz+vj76WDO+jB/zy+B3UYEJmKwnswjkIAJ/N5kQNTDK+m1yvvT2n+cvf 981UwAxaLeFGQrCfJdafTxxacWwI1dK9Q2YqoCMFfgSwBxA9+WDzCRAD4OWx24e6A8dCSdrt0ra4M2KP Dx1N51r1UNk231QgCLO6Evsyk0+GshA9OBHLZwgCb+pOYiKwkQ8If1/nVBANLeAOwMlEqCF59aoDrTKA ng0JCLTy2Aqb1LFQlMDXTA9Ftr1HtdKiAOIuB5AJv0NykBSRG4qgP8uZlXl5fVgAfw+WhsDnyNADjoKf ipmY72niXwj2T8gG8DfTzgtxAANv+YeggmHHsNSOOQy/vXOvnHtCIRAIB9b8D4+Dhks1lrLoCSAN7j63 y0OC0jYk5AAm/cKsCnEuACP1UCvglDUQ0YPLsT8Bp4+4Pgx375sJYUGSzuVHF8HJ7fhwSMp/E4SIB2/L lIIKAGiPe/nN+qZf5N3r9QKKjGn7oigGIx2BNgUgE4IRhJAAmgsH2hIgw+WBTbiflQEZPFrQJ45j9MCO AzV8DVL3D3wGo4399qBLoR8BL4K1l1o+cPY2xffi22bNFmaGl5JtDeG4UIage/BwlUvD833+5CuuNv85 omL+9fLE7HlvzDFZkAAOw9AaOjo0YVcPGpBVDYvlAHf2X7qmm68I1dK7S9BHF7e1sOgIJfUgJhVICNDF zVg9v7VkFfZpMb8Kn+SljQXwU/HaZBhmSI9zZD8lCfXZu3R+vLbIL7h9ZHAn8UEihn+18E63l8jpN6tZ 7/MGRAmoZGMl3K+9ti/7iTf7hqIgBXMnB0dNRYEaAKgBpVAZwEsGKAYUSSBCCpAZvkd5GASwWEIYOpgY 5yydDk6fmVS/25kv1C9WLZos2hJH+YMMDk+eXDOMMRgQv8JjKg3v/R8EZoaXlGSX/f2L/uCMA0KAQrAp IKwFDgYnuTRgB0iCjPCUwNdGj5gK87lyQKeFvGP6znr3XEmIkINj7fK3v7VL9+bwI+f85FFBhyRAC7ep 4k+mr1+j7xf3TwyyTw8EyvvPPPhwjIlmSa+HPF/sVi/N4foEYCAKhNBdBQgBomBCUSuLFrhdZIlEQoIH UE+mb++TVsPiAqEZRP42Xg52YDuOvedqUhAW0s4tfK/SvptTA10BEb+H2z/TDRpg7hdJ7GayIBtu2Xmt YRaAE/jvqiib+58P4AMRJAVBWAlQBuEllgbwDOu8MwIukcAI//qYf3UQIS6OM6IozbSKYyDddEAi4i8F ECPgQhtBfHDfxQsn+ijYBePoHXSwlc3anlEXyJgD7/4GS3lviTev6TqvvzVTMBALhLgkNDp8U9ApwEpv YuVVeuAnDCEDYL0a7CJBOCpopA1Nh/Ns4GxNHmfZlNsKplwE4AYb2977Vyv2zRZujLbFLZ/STA7yIBPI ZLO3HXSwUEScAGfF9SeCW9Vo36cvX8J+n9ARIgAE4CyGiHDx82lgWx3bewfSH85Y0ligS4aqBdgZgPwC GXSeYA+L1P3O/TBxA3EUiHf2KI0JfZpCsDA2DjuK5qGYCNz/fC5fxWuHtgtRW0cWb+AyTwQXkOPz9xNz IJCHF8mMdY829peaYuvD9ATAQA4FYBo6OjkE6njaEAJgUR/KgGbJuK8JCLuFVAVK9vywMkoQJcJGC6jm S6dELwIQPbfaofNj7fC32ZTTDct0WBGomIAz3pEKB0vItI/uCJu5wMvKxS9qvFYCwdkP6m036STPzRlQ gBFItyLmDoWN6YEKRKAAlgau/SQFUAZwvQvQSoHpLKAZhyAlGVQFwkEGU6sWRTAx1wvr8V+jKbVA7BZi OZLtidPQI3dzTD5fzWAMhN17CePyxRqOYeTfLbwe+XF+jXdw1G8PxU+vPtvknv+LOt2AgAwE8F2BKCNN af2rsU/vLGEjEfQBUA7i0whQI/ZstbisMSwI/ZBQq4pg7AsCSQVBgQxvvbiMDHTJ7dF/xhgG8jEJP3p+ fsScdtB4/j9lACNXh/GCsfBz410KFJf2PTzyx6f4AECaBYNKsAW0KQgvwvbyyB8/2tgXwAbh+mewkwHJ BCAfwaejqxr/yXZgP4hgS2MCAJ+R8F/C5C4IC/e2B1aBKwgZ+HBtL39vH+MNGmHa3lIgHu+W05ASSZqJ 7/zqluBX7c7Wcq+01OTmrDPh8rAgAwzwrgKsCWEKRARyXA8wHYQORLAmFDBRfgbeGAVO+fjRxAlGsYzx 8F/D45AF+VYfL+d6+lRPCbSIA+L1UHNDPU/H2z/tjwg1t9aexvqvkXi7Pj/QESJgBTRWBo6LQzIciTgx IJoOznLcVSh+CNXSu0M/F8iMAE/HrOAfheEdC+3l8CdNQwgL7f9jO4PgMm2nTwG0jApgJMSgDgRYBvVk f2/PRsP571tyX+ki778RU7AQD4dQcODZ2GzWvcKkDtC2hvgr+8sUTLB2AowJuIEOg8pkfw012IUljgC3 xXLqCecwBhlIAE+LCenr/X9HP7hgF3r6XK4P97UyQSCD4fJIFaM/63Bttl8Fs6/orF5BN/dCVKAKaBIa gCXL0BptZhSgKmTkIkBx8SKGxfaM0P2EIA6TWT508K/GGUgI/nl6S/j3d2EYSNsHxUQOl4F/z1bBnsCH oK/rvXKuBGAvDIB1SBr5MAz/qHVQJ3TpW7/WjWv14Sf3QlQgAAbhWAJJDJZLxDASr96Xts7cS8AkBBz6 9oNnBLSUFXLmC2VEASnh9JI4zcp+/Bfy8aPr5/aL26+qiAmXdXKq/PjRNBmHxAUA28qA7piBrzY8lPiv ttNf9icXa9P8AsEACqgGJRP0kIJdDQsTy0tDwTigBsnYS0mehie5NIAibg8w1JqCBMMb8pYWiaFsRtLs OBsErAFhZIkl4ixps7mrV/uy/4fYCP4UCAADxzAsr7k7jfF/DcaNyvzfgj7b58xn+xOPveHyBBAgCQy4 KTk5MBFXD48GFng5BvspDuKaDzBZAEfswuUGcOmMBPtyXjaHJXOOAKBeh9ksCPqgh8AW+K49Hwd4VHwC Po8SopARNB/fVsGeh49SYCixLQM/9MCbDTeaMoAKz3o/QfeOMd81Zfg/R/4ghACwWuXQmGAsfykM1mQ+ UDbCRg6iSkOQE+kchkNMxwlQtN+wXmAvwuQgibA3AZgp5f7x4ozzXEK6oD08818+5K+OvZlDf4AyQgJQ U9S4NRwE/fg/V+Gvcbe/1Z1h8P+Lx3716SkAysRAkAwG+CMFYFwpQGXYlCSgI0aYgk8GNWnkVgIwGqBl zAx+clNTCXJJCE3dzRDDd3NMONXSucCgDJgCsACnwTAYRRAqJZSMCU9Atjm9c0yXG/I+s/F9IfV+IEAO AeIEpDgc1rmmomAaoEsJ2YJw2xRdgEehpG0FACk5C2SgEnBAn0TwIJINCpIQkg+GdeXa5AjyQw8+pymB roMAKfg99HBSD4nTkBiQQmwif9uGHSj0p/W6//XEt/XLNKALYOQU4CtRIA7R0QOwkr3twVAkj5BKomfJ RAEqVAUww+V8DnBMDDAAp+PAFaAn6tZGANBSQVMFEb+HGwZ9i4f7YbfkxrVggAwN4hyKsCuVwulqQgzg 9AEuA7CxHEfDQZHUzCn5M2JpmShLx6gP0H+PVz7cGjAh//DXjlhl9HPT8m/6jHN3n/WpSAszpAKgMw0a b1+Uex8/2tkFrYEoj7pZLfXDb8mNasEQCAfygwNHQa+jKbYiGBzxZU9w3wTkJKFCjtTSTgmlZEiUCaKY h2c0ezSCKUHCh46sGwakLNRgCUCDAM+8sbS+CvZ1OBqw8J1BISmJSAb7nPZjzjL4G/HuN+uuaEAHxDgU wmE0tlgCcHbTkGqgZcJGA680DafkyVgYkIXD//bCoHlPLYVk33UdhI4MauFRroOfhn3l0ZAP+skUDF+9 +9loIHJ7tjA7803sun5DfvCADAr00YqwIDb7wD6XQ6VhL48rlFMLV3qfNr+ZxCOqqMTiyykcnXnUsCLc ZcHeAmpbD/FlQNcZECBTw1H/Dz3wsHPxoFfxQSCFsVkEiAxvxR5T+W+zDpxyf72sBfLNaH9Mc16wQAYJ 8bwPMBSAJxVAYoEWDs7/panFBESQArC1KZ0fb9pG3KVBkg2KL+u+jR6xywHLgIcPzZ6XHtklEikBQSBb +NBGpVArWUBmGiDe4fWh8AdBgiwB5/LPe5+vzrMe6na04JwGevAJJAS8szsZIAAtP36/gfOv+jlxKMps /6unNJIEzgygC9sc/PKA1PpVe+eYqqIU4cnAToxilblQR/F+f7W43gjyMXEKVJ6O61cp2fAj2K9zeB3z fpN9clP2nNCQEAmEMBEwnkcrlESCAsYZg83vn+VnV+YdjPRLDZTj5GINrUBgc+vdqATw1JACsePt2SnA RwklNUEgjTG+BDBnGD35XxRwLA1vdisb7ifrrmjAAA/PIBNCmYlBIITQTtTarHQJK9tch4DmZMNuJnUv Di+Ym+n4VfjyRAz1lEw0NaOQmYdlyi8RyJKxRw5QTiUgJ0ZiAHvy8ROMHPNvnUe9xP15wTgCsfQJOChw 8fhlwuF1u3YBwgxaQiVQIIhjBZ/tn4WTmpSCRAwwAfEpB6JVxKAIEv5QHwcRTw03ta45fA70sCtYC/Xp p9bGtOCQBAzgdIY8SQBOolHJAARtuP41QDcZOASwV83blETVv2DQds1ZKoKsAVCpiIgEp+17VW8Pts8q lX8APUAQEAWPoDJBI4lldKIM4SYZxGjz5XexHa6+Pn9FUCUXMClAgkZeRbEfBRApQI8F4aGsrBnhT4H4 ekH191QQAA/k1ClAQwJ1CPJKAAVznE1LSfYLbBbyMBCfimMMClBOjVtzRIScC3NEhJwDY5OEwOAJt8JP C7av31nvTjq24IAMC8X0CsDFQahbBPIK624aRtrvMCroSglBMwJQZ9wwEEt6tM+NezKZjauzS0ErCdFx A2BzA10AGphS1Gz+9T66fgr/dVdwRgSgqayoM0HHhcSKAejBIAElOYMABzBL6lQd9cgFQmtCkB1+hw3z DgwclubWNPHOCvZ8+Pq64IACAcCfCcQCaTqZsKweNgvklBSgI0MRgmFDCVTH17A7gSsE0OjqIERjJdTv DzbD8F/+OQ8ZdW3REAgL0ywM8XQBLA6kAmk6m7CkE9my8J0DCg1sqAjxKQqgNICD4jyn3A/+BkN5SOd8 Er6bXaxp4wnr8eN/iEWXVJAADu8qAtMZjNZus+OVgvZiMAbECiRIAKIIoSoIQQRQncPeB3ipBNBVASuD XYDpvXNFnBb8v2Pw6NPq5VtwQA4K8EpGYhTA42QoJwRPDZAns+AGN/mgNw5QNqyQnMvLtSTVFyDTV1kQ AFP47upsk+n2z/kwR+gDonAIDaSaAREviBn5KAKx+ACgCBj/e+4EcV4NovgOCnh4iYSCBMMhAlvwn888 Hz46p7AgCITgJIBNg52KgSuInAVwlIRODbJOTaPsxPEkKrVQFMDXSIkn++gh/gMSEAgGgkgFuJ8T83nU 431EAIEvCpDPBj2m1KwJQQPN/fClN7l8LtfavUKDQEP04PrpUEcHAnen0KfvwbwTySLdtfLD6+CT9pPT YEAOBHAqZeAUoEqAYaRBAEvy8JmBqEfJUAVQP0HAFphPjdA6th+q01kQjg/qH1Ackvgd+0pZeDv1h8Mj w/rseKAADCkwDPC3A10KgUyETAycAWBlAlwKsDNhK4sWuFGkVGpwff3NGsThBC4PuoAJ4HoIk+OsTDVO az7eorFp8sz4/rsSMAXJQEpA1EppCAE0FLyzOweU1TgwgsJEBnD/iQgK0/wHaWAJIAP0x05tXlWiLQpQ RorC95fSnen4/gB3iMCYB3DLp6BaT2YZR/WCmI40CSJ8nCbB92kYDPGQLo+W/sWqFyAPgcgh/JwAR+7O ijwJc29FiTffME/ACPMQEAVEmgVCqJJOAKCUxhQSM/EI4EMBeA1QA+lNR1hgAlAX6uIBICJgZNpwmj3P fx+ij55zv4AR5zAgDQlUDYvICkBrLZbIMIGAF8tiAYBkhJQT5INAwJ0LME6aGi6PG5EsDr+f5W2LymKe D1pcYem+THGX7SMI8nFfwATwAB4OIkQPMCk5OT5pCAqAFaKWgQQZAEuBLgwLedIxCGBGg+AMHPS4M3dq 0Q4/y4vP7jMMwjjvXEEACATALFoq4GJr741qoGOAlQIpjPbcV0mjCfG2g6RyAMAbgSgQj+kUyXN/DR61 tLfA7J/ySDH+AJIwAAj5DAkCDkuQFKBJgoxGRhamELnGt98smAVwIQ/PjYRABhVQCCn54szI8X9wE+r+ vToZ2m+v58ivel9cQRAC4fNSDlBujOQik/kM1mFRFQVfAkkIGpFZgSAEp/rgRcKgCBbwsBsC8AE4FTAx 2wZ90SMbnHgU/J2hXr2yT/fAI/wBNMAAAeJOBKEloUARIBhgePc66AN/5wwEtXVxjAVQAefOKTC0Bvj4 k9F/C1br4K8Gmsb/P6xeL0vJL8fD3RBAAghwQ+akDKD/CyIVcF6XQaUgtbILWwRZFBvRMCB75ECibw08 c+KsAE/us9y2Ak0wV71i2B1MIWzdtLmX1bnG8EPvP6xeL8i/el9cQTAC5KAqbcgJMIHIoAiYAqg9TCFt VpWK9kwEmAA990LzUChckF7Fm3RHl62rJr8vaSx3fJfZfXn2+Sn695QwAAnmoggiKQKgf4x8uVAQ0V6o kQpJHhEhlIyUCfisDF9iZN2ts8vc3bS1KfAr/h9cOteUUAuEyVgmIxvCKIQgaoDGi4gAphrkhBCgV8FQ CfD3Cutax4UNJLgJfM5u15I48P8E0Z/vnu9emalwQAECSBqEQgqQL6R8vJwBYqcFKYLWKQwM89P+0DQL BToFPP7gt4WybfJPOjAL/h9c1r3hIALhsR4LhniQjoZGKXKuBk4CIEEylwckCCQLvYXiULvHcBPzAItL 1J+0z8PhLAJbCHAbwk7ynoeXwfBvjFYsPr+6x5TwC4bPmBYtGsCKyqwEMZIBmYQgaRGBZ3WgkiCcNSpy /QTWCXQM8n8uDvjYIer1pyLwTwG+CXV4MA2ApDBBNffBsIDy5duqT90ZrIIAwhcKXAyYETBO1N4IbkYT L8HB+Am4DuA3atfFc514F6epO3dwG/IffDrQYBGJYPEfiqAhsZ2JKIYYjBRBKcLEyPbVeTcZCbwC7Keh bTc08vgr7h8WNfDQJwLFOOwJgnqOw1oINKeZehlECUFALvNzApBhNJULKQiIM/Z3q/9H0kj849OyU47d 9JEnk2Ty95+wbw410NAvBcLiIoFmVVwJWBTR1IKkFSCkgMNtVge87nNZMSwe/LgR7w7B4enpKjCfQumd 8Afm2rQQAhl0QEPmTRpst5AAABr0lEQVSAMwmkUEFKJHJiUKRAwCURg80k0ErPSR4cvbhWkiNxuxS/+4 D90qVLKpfCQd8AfvKrQQA1LJsqsCoDMpvARyFwUpDIIUASLLSgisL4GvPc1INzgJtA7uPdTV7eB/QN4M e7GgQQw6KzCX3JIEAIQv5AIoYwBGEiDBeQTZ/hC/IA0FkCz+Tli8UG6Gd7NQgg5uVDBmEIwaYUTMaByc EqEYn0mvSZJpBLmXqXhy8Wp7XfiQn0DeAntxoEkOAykYGNEIpFD1IwkAMliTCEIYGa3qsY3QJwCegc7M WiGfD4O2qAfnZXgwBmaXEycCkETgrF4nQAXJOTk05Axm3Sz8B/zmJxOvBv4QTY8PL1sRoEMEfLhxBMSk EiBzQEpATUsEY/jxv9WXzB3gB8/a0GAdTRouCQwgYfgrARhs/VZfR7+wK9Afj6XQ0CqPMlAUlSDGGMA9 gH0BLAG0B//FeDAB7jJYHPRBQcsC4Q26yxnpzVIIDGaqx5vBoE0FiNNY9XgwAaq7Hm8WoQQGM11jxeDQ JorMaax6tBAI3VWPN4NQigsRprHq8GATRWY83j9f8BXw26zxg/U0gAAAAASUVORK5CYII=" width="2 56" height="256"><br>
354 <div id="cpu">CPU 1x Gaussian blur</div>
355 <div id="cpustat"></div>
356 <canvas id="2d" width="256" height="256"></canvas>
357 <div id="gpu">GPU 1000x Gaussian blur</div>
358 <div id="gpustat"></div>
359 <canvas id="gl" width="256" height="256"></canvas><br>
360 </body></html>
OLDNEW
« no previous file with comments | « conformance/more/index.html ('k') | conformance/more/performance/bandwidth.html » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698